Skip to content

Instantly share code, notes, and snippets.

@XinyueZ

XinyueZ/laptop_price_prediction.ipynb

Created July 29, 2022 09:39
Show Gist options
  • Save XinyueZ/6b27c0d4d77f1fb83cd127a178dee908 to your computer and use it in GitHub Desktop.
Save XinyueZ/6b27c0d4d77f1fb83cd127a178dee908 to your computer and use it in GitHub Desktop.
Download ZIP
laptop_price_prediction.ipynb
Display the source blob
Display the rendered blob
Raw
laptop_price_prediction.ipynb
{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "view-in-github",
"colab_type": "text"
},
"source": [
"<a href=\"https://colab.research.google.com/gist/XinyueZ/6b27c0d4d77f1fb83cd127a178dee908/laptop_price_prediction.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "dGVlhVSzjyp6"
},
"source": [
"Prediction of laptop price\n",
"====\n",
"\n",
"Continue the previous notebook and use the cleaned data to process regression.\n",
"\n",
"## Goal\n",
"\n",
"- Try different regressors and figure out the differences between them.\n",
"\n",
"- Visualize the result of regressions.\n",
"\n",
"- Ensemble several models into a best one.\n",
"\n",
"- Build Deep-learning models to compare others."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "zxK7aZNfjwbL"
},
"source": [
"Notebook clean data: https://colab.research.google.com/drive/1ZRtM_-y3nS_gZKMzRyClU51qQ16B28KY?usp=sharing\n",
"\n",
"\n",
"Notebook share: https://colab.research.google.com/drive/1w_IhEC5Y41zCQ_ODFeGPB1zR9BfjyrvM?usp=sharing\n",
"\n",
"Report https://www.linkedin.com/posts/chris-xinyue-z-547b5486_machine-learning-deep-learning-activity-6958555088478998528-kJ5V"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "vbdSiAwFLOVf"
},
"source": [
"## Imprint\n",
"\n",
"Contact: chris.at.de@gmail.com\n",
"\n",
"\n",
"#### About me \n",
"\n",
"I am an ML advocate and enthusiastic developer. Please share this notebook for growing skill purpose.\n",
"\n",
"When you are so interesting in ML and AI like me, please conntact me, let's exchange knowledge."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "pLv5BvqqjXBi",
"outputId": "0cb9bff6-2d0d-4eec-bb58-78bd70de9681"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
"Requirement already satisfied: gdown in /usr/local/lib/python3.7/dist-packages (4.4.0)\n",
"Requirement already satisfied: six in /usr/local/lib/python3.7/dist-packages (from gdown) (1.15.0)\n",
"Requirement already satisfied: tqdm in /usr/local/lib/python3.7/dist-packages (from gdown) (4.64.0)\n",
"Requirement already satisfied: beautifulsoup4 in /usr/local/lib/python3.7/dist-packages (from gdown) (4.6.3)\n",
"Requirement already satisfied: requests[socks] in /usr/local/lib/python3.7/dist-packages (from gdown) (2.23.0)\n",
"Requirement already satisfied: filelock in /usr/local/lib/python3.7/dist-packages (from gdown) (3.7.1)\n",
"Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests[socks]->gdown) (1.24.3)\n",
"Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests[socks]->gdown) (2022.6.15)\n",
"Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests[socks]->gdown) (2.10)\n",
"Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests[socks]->gdown) (3.0.4)\n",
"Requirement already satisfied: PySocks!=1.5.7,>=1.5.6 in /usr/local/lib/python3.7/dist-packages (from requests[socks]->gdown) (1.7.1)\n",
"/usr/local/lib/python3.7/dist-packages/gdown/cli.py:131: FutureWarning: Option `--id` was deprecated in version 4.3.1 and will be removed in 5.0. You don't need to pass it anymore to use a file ID.\n",
" category=FutureWarning,\n",
"Downloading...\n",
"From: https://drive.google.com/uc?id=1NjeScECu01RCIqp21Wxw0MWawE0oPI3d\n",
"To: /content/laptop_price_cleaned.csv\n",
"100% 81.4k/81.4k [00:00<00:00, 71.9MB/s]\n"
]
}
],
"source": [
"!pip install gdown\n",
"!gdown --id 1NjeScECu01RCIqp21Wxw0MWawE0oPI3d -O ./laptop_price_cleaned.csv\n",
"\n",
"csv_file = \"./laptop_price_cleaned.csv\""
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "EWHFUH_MjgqT"
},
"outputs": [],
"source": [
"import re\n",
"import numpy as np\n",
"import pandas as pd\n",
"import seaborn as sns\n",
"\n",
"from sklearn import preprocessing\n",
"from sklearn.compose import ColumnTransformer\n",
"from sklearn.model_selection import GridSearchCV\n",
"from sklearn.model_selection import train_test_split\n",
"from sklearn.metrics import mean_squared_error, r2_score\n",
"\n",
"from sklearn.tree import DecisionTreeRegressor\n",
"from sklearn.neighbors import KNeighborsRegressor\n",
"from sklearn.linear_model import LinearRegression, Lasso, Ridge, ElasticNet\n",
"from sklearn.ensemble import BaggingRegressor, RandomForestRegressor, GradientBoostingRegressor, AdaBoostRegressor, VotingRegressor\n",
"\n",
"\n",
"import matplotlib.pyplot as plt\n",
"import matplotlib.colors as mcolors\n",
"\n",
"import tensorflow as tf\n",
"import tensorflow.keras as keras\n",
"\n",
"import time\n",
"from dateutil.relativedelta import relativedelta as rd\n",
"%matplotlib inline\n",
"\n",
"np.random.seed(42)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "0hw4VQ2ev08B"
},
"source": [
"## Data class\n",
"\n",
"- Process data\n",
"- Display data meta\n",
"- Clean or norm data"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "BNmdWCzEl-S5"
},
"outputs": [],
"source": [
"class Data:\n",
" def __init__(self, src) -> None:\n",
" self.data = pd.read_csv(csv_file)\n",
" self.data.drop('Unnamed: 0', 1, inplace=True)\n",
"\n",
" self.transformer = preprocessing.StandardScaler()\n",
"\n",
" def display(self):\n",
" display(self.data.head())\n",
" display(self.data.describe().T)\n",
"\n",
" def norm(self, Y_name):\n",
" fields = list(self.data.columns)\n",
" fields.remove(Y_name) \n",
" self.data[fields] = self.transformer.fit_transform(self.data[fields])\n",
" \n",
" def plot_corr_with_Y(self, Y_name):\n",
" y = self.data[Y_name]\n",
"\n",
" fields = list(self.data.columns)\n",
" fields.remove(Y_name)\n",
"\n",
" correlations = self.data[fields].corrwith(y)\n",
" correlations.sort_values(inplace=True)\n",
"\n",
" display(correlations.sort_values().to_frame().T)\n",
" \n",
" colors = list(map(lambda dt: \"red\" if dt <= 0 else \"green\", correlations))\n",
"\n",
" correlations.plot(kind='bar', color=colors, figsize=(10, 5), fontsize=12)\n",
" \n",
" plt.title(\"Correlations to price\")\n",
" plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "SYshrxVDvxqy"
},
"source": [
"### Display data"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
},
"id": "gLqkE3XYmvzx",
"outputId": "6dad46b8-c3c0-40be-e83b-c511f8d7ca5d"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:4: FutureWarning: In a future version of pandas all arguments of DataFrame.drop except for the argument 'labels' will be keyword-only\n",
" after removing the cwd from sys.path.\n"
]
},
{
"data": {
"text/html": [
"\n",
" <div id=\"df-18cbbe62-34d6-4738-bb5e-25d4ae40a9b2\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Company</th>\n",
" <th>Is_MacBook</th>\n",
" <th>Type</th>\n",
" <th>Inches</th>\n",
" <th>ScreenResolution</th>\n",
" <th>Ram</th>\n",
" <th>Gpu</th>\n",
" <th>OpSys</th>\n",
" <th>Weight_kg</th>\n",
" <th>Price_euros</th>\n",
" <th>Cpu_GHz</th>\n",
" <th>Cpu_Brand</th>\n",
" <th>Cpu_Core</th>\n",
" <th>Storage_GB</th>\n",
" <th>Storage_Type</th>\n",
" <th>Screen_Width</th>\n",
" <th>Screen_Height</th>\n",
" <th>Is_Touchscreen</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>13.3</td>\n",
" <td>0</td>\n",
" <td>8</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1.37</td>\n",
" <td>1339.69</td>\n",
" <td>2.3</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>128.0</td>\n",
" <td>0</td>\n",
" <td>2560</td>\n",
" <td>1600</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>13.3</td>\n",
" <td>1</td>\n",
" <td>8</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1.34</td>\n",
" <td>898.94</td>\n",
" <td>1.8</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>128.0</td>\n",
" <td>1</td>\n",
" <td>1440</td>\n",
" <td>900</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>15.6</td>\n",
" <td>2</td>\n",
" <td>8</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>1.86</td>\n",
" <td>575.00</td>\n",
" <td>2.5</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>256.0</td>\n",
" <td>0</td>\n",
" <td>1920</td>\n",
" <td>1080</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>15.4</td>\n",
" <td>0</td>\n",
" <td>16</td>\n",
" <td>2</td>\n",
" <td>0</td>\n",
" <td>1.83</td>\n",
" <td>2537.45</td>\n",
" <td>2.7</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>512.0</td>\n",
" <td>0</td>\n",
" <td>2880</td>\n",
" <td>1800</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>13.3</td>\n",
" <td>0</td>\n",
" <td>8</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1.37</td>\n",
" <td>1803.60</td>\n",
" <td>3.1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>256.0</td>\n",
" <td>0</td>\n",
" <td>2560</td>\n",
" <td>1600</td>\n",
" <td>0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-18cbbe62-34d6-4738-bb5e-25d4ae40a9b2')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-18cbbe62-34d6-4738-bb5e-25d4ae40a9b2 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-18cbbe62-34d6-4738-bb5e-25d4ae40a9b2');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" Company Is_MacBook Type Inches ScreenResolution Ram Gpu OpSys \\\n",
"0 0 1 0 13.3 0 8 0 0 \n",
"1 0 0 0 13.3 1 8 1 0 \n",
"2 1 0 1 15.6 2 8 1 1 \n",
"3 0 1 0 15.4 0 16 2 0 \n",
"4 0 1 0 13.3 0 8 0 0 \n",
"\n",
" Weight_kg Price_euros Cpu_GHz Cpu_Brand Cpu_Core Storage_GB \\\n",
"0 1.37 1339.69 2.3 0 0 128.0 \n",
"1 1.34 898.94 1.8 0 0 128.0 \n",
"2 1.86 575.00 2.5 0 0 256.0 \n",
"3 1.83 2537.45 2.7 0 1 512.0 \n",
"4 1.37 1803.60 3.1 0 0 256.0 \n",
"\n",
" Storage_Type Screen_Width Screen_Height Is_Touchscreen \n",
"0 0 2560 1600 0 \n",
"1 1 1440 900 0 \n",
"2 0 1920 1080 0 \n",
"3 0 2880 1800 0 \n",
"4 0 2560 1600 0 "
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <div id=\"df-bd55a218-794c-471f-a77d-b97006ec2c45\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>count</th>\n",
" <th>mean</th>\n",
" <th>std</th>\n",
" <th>min</th>\n",
" <th>25%</th>\n",
" <th>50%</th>\n",
" <th>75%</th>\n",
" <th>max</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>Company</th>\n",
" <td>1303.0</td>\n",
" <td>3.907905</td>\n",
" <td>2.751755</td>\n",
" <td>0.00</td>\n",
" <td>2.0</td>\n",
" <td>4.00</td>\n",
" <td>5.00</td>\n",
" <td>18.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Is_MacBook</th>\n",
" <td>1303.0</td>\n",
" <td>0.014582</td>\n",
" <td>0.119917</td>\n",
" <td>0.00</td>\n",
" <td>0.0</td>\n",
" <td>0.00</td>\n",
" <td>0.00</td>\n",
" <td>1.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Type</th>\n",
" <td>1303.0</td>\n",
" <td>1.551036</td>\n",
" <td>1.299402</td>\n",
" <td>0.00</td>\n",
" <td>1.0</td>\n",
" <td>1.00</td>\n",
" <td>3.00</td>\n",
" <td>5.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Inches</th>\n",
" <td>1303.0</td>\n",
" <td>15.017191</td>\n",
" <td>1.426304</td>\n",
" <td>10.10</td>\n",
" <td>14.0</td>\n",
" <td>15.60</td>\n",
" <td>15.60</td>\n",
" <td>18.4</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ScreenResolution</th>\n",
" <td>1303.0</td>\n",
" <td>1.778972</td>\n",
" <td>0.665327</td>\n",
" <td>0.00</td>\n",
" <td>1.0</td>\n",
" <td>2.00</td>\n",
" <td>2.00</td>\n",
" <td>4.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Ram</th>\n",
" <td>1303.0</td>\n",
" <td>8.382195</td>\n",
" <td>5.084665</td>\n",
" <td>2.00</td>\n",
" <td>4.0</td>\n",
" <td>8.00</td>\n",
" <td>8.00</td>\n",
" <td>64.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Gpu</th>\n",
" <td>1303.0</td>\n",
" <td>2.036071</td>\n",
" <td>1.367638</td>\n",
" <td>0.00</td>\n",
" <td>1.0</td>\n",
" <td>1.00</td>\n",
" <td>3.00</td>\n",
" <td>10.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>OpSys</th>\n",
" <td>1303.0</td>\n",
" <td>2.029931</td>\n",
" <td>0.595576</td>\n",
" <td>0.00</td>\n",
" <td>2.0</td>\n",
" <td>2.00</td>\n",
" <td>2.00</td>\n",
" <td>5.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Weight_kg</th>\n",
" <td>1303.0</td>\n",
" <td>2.038734</td>\n",
" <td>0.665475</td>\n",
" <td>0.69</td>\n",
" <td>1.5</td>\n",
" <td>2.04</td>\n",
" <td>2.30</td>\n",
" <td>4.7</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Price_euros</th>\n",
" <td>1303.0</td>\n",
" <td>1123.686992</td>\n",
" <td>699.009043</td>\n",
" <td>174.00</td>\n",
" <td>599.0</td>\n",
" <td>977.00</td>\n",
" <td>1487.88</td>\n",
" <td>6099.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Cpu_GHz</th>\n",
" <td>1303.0</td>\n",
" <td>2.298772</td>\n",
" <td>0.506340</td>\n",
" <td>0.90</td>\n",
" <td>2.0</td>\n",
" <td>2.50</td>\n",
" <td>2.70</td>\n",
" <td>3.6</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Cpu_Brand</th>\n",
" <td>1303.0</td>\n",
" <td>0.049117</td>\n",
" <td>0.219720</td>\n",
" <td>0.00</td>\n",
" <td>0.0</td>\n",
" <td>0.00</td>\n",
" <td>0.00</td>\n",
" <td>2.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Cpu_Core</th>\n",
" <td>1303.0</td>\n",
" <td>1.770530</td>\n",
" <td>2.375694</td>\n",
" <td>0.00</td>\n",
" <td>0.0</td>\n",
" <td>1.00</td>\n",
" <td>3.00</td>\n",
" <td>12.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Storage_GB</th>\n",
" <td>1303.0</td>\n",
" <td>606.851880</td>\n",
" <td>463.309442</td>\n",
" <td>8.00</td>\n",
" <td>256.0</td>\n",
" <td>500.00</td>\n",
" <td>1000.00</td>\n",
" <td>2512.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Storage_Type</th>\n",
" <td>1303.0</td>\n",
" <td>1.136608</td>\n",
" <td>1.223077</td>\n",
" <td>0.00</td>\n",
" <td>0.0</td>\n",
" <td>1.00</td>\n",
" <td>2.00</td>\n",
" <td>6.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Screen_Width</th>\n",
" <td>1303.0</td>\n",
" <td>1894.784344</td>\n",
" <td>494.641028</td>\n",
" <td>1366.00</td>\n",
" <td>1600.0</td>\n",
" <td>1920.00</td>\n",
" <td>1920.00</td>\n",
" <td>3840.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Screen_Height</th>\n",
" <td>1303.0</td>\n",
" <td>1070.830391</td>\n",
" <td>284.519410</td>\n",
" <td>768.00</td>\n",
" <td>900.0</td>\n",
" <td>1080.00</td>\n",
" <td>1080.00</td>\n",
" <td>2160.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Is_Touchscreen</th>\n",
" <td>1303.0</td>\n",
" <td>0.147352</td>\n",
" <td>0.354593</td>\n",
" <td>0.00</td>\n",
" <td>0.0</td>\n",
" <td>0.00</td>\n",
" <td>0.00</td>\n",
" <td>1.0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-bd55a218-794c-471f-a77d-b97006ec2c45')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-bd55a218-794c-471f-a77d-b97006ec2c45 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-bd55a218-794c-471f-a77d-b97006ec2c45');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" count mean std min 25% 50% \\\n",
"Company 1303.0 3.907905 2.751755 0.00 2.0 4.00 \n",
"Is_MacBook 1303.0 0.014582 0.119917 0.00 0.0 0.00 \n",
"Type 1303.0 1.551036 1.299402 0.00 1.0 1.00 \n",
"Inches 1303.0 15.017191 1.426304 10.10 14.0 15.60 \n",
"ScreenResolution 1303.0 1.778972 0.665327 0.00 1.0 2.00 \n",
"Ram 1303.0 8.382195 5.084665 2.00 4.0 8.00 \n",
"Gpu 1303.0 2.036071 1.367638 0.00 1.0 1.00 \n",
"OpSys 1303.0 2.029931 0.595576 0.00 2.0 2.00 \n",
"Weight_kg 1303.0 2.038734 0.665475 0.69 1.5 2.04 \n",
"Price_euros 1303.0 1123.686992 699.009043 174.00 599.0 977.00 \n",
"Cpu_GHz 1303.0 2.298772 0.506340 0.90 2.0 2.50 \n",
"Cpu_Brand 1303.0 0.049117 0.219720 0.00 0.0 0.00 \n",
"Cpu_Core 1303.0 1.770530 2.375694 0.00 0.0 1.00 \n",
"Storage_GB 1303.0 606.851880 463.309442 8.00 256.0 500.00 \n",
"Storage_Type 1303.0 1.136608 1.223077 0.00 0.0 1.00 \n",
"Screen_Width 1303.0 1894.784344 494.641028 1366.00 1600.0 1920.00 \n",
"Screen_Height 1303.0 1070.830391 284.519410 768.00 900.0 1080.00 \n",
"Is_Touchscreen 1303.0 0.147352 0.354593 0.00 0.0 0.00 \n",
"\n",
" 75% max \n",
"Company 5.00 18.0 \n",
"Is_MacBook 0.00 1.0 \n",
"Type 3.00 5.0 \n",
"Inches 15.60 18.4 \n",
"ScreenResolution 2.00 4.0 \n",
"Ram 8.00 64.0 \n",
"Gpu 3.00 10.0 \n",
"OpSys 2.00 5.0 \n",
"Weight_kg 2.30 4.7 \n",
"Price_euros 1487.88 6099.0 \n",
"Cpu_GHz 2.70 3.6 \n",
"Cpu_Brand 0.00 2.0 \n",
"Cpu_Core 3.00 12.0 \n",
"Storage_GB 1000.00 2512.0 \n",
"Storage_Type 2.00 6.0 \n",
"Screen_Width 1920.00 3840.0 \n",
"Screen_Height 1080.00 2160.0 \n",
"Is_Touchscreen 0.00 1.0 "
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <div id=\"df-9fa09613-00a3-4f41-b6a3-69361373fa28\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Cpu_Core</th>\n",
" <th>Cpu_Brand</th>\n",
" <th>OpSys</th>\n",
" <th>Storage_Type</th>\n",
" <th>Inches</th>\n",
" <th>Is_MacBook</th>\n",
" <th>Company</th>\n",
" <th>Storage_GB</th>\n",
" <th>Is_Touchscreen</th>\n",
" <th>Weight_kg</th>\n",
" <th>Type</th>\n",
" <th>Gpu</th>\n",
" <th>Cpu_GHz</th>\n",
" <th>ScreenResolution</th>\n",
" <th>Screen_Height</th>\n",
" <th>Screen_Width</th>\n",
" <th>Ram</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>-0.392649</td>\n",
" <td>-0.179219</td>\n",
" <td>-0.117135</td>\n",
" <td>-0.079704</td>\n",
" <td>0.068197</td>\n",
" <td>0.0865</td>\n",
" <td>0.149303</td>\n",
" <td>0.149772</td>\n",
" <td>0.191226</td>\n",
" <td>0.21037</td>\n",
" <td>0.292191</td>\n",
" <td>0.333036</td>\n",
" <td>0.430293</td>\n",
" <td>0.453002</td>\n",
" <td>0.552809</td>\n",
" <td>0.556529</td>\n",
" <td>0.743007</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9fa09613-00a3-4f41-b6a3-69361373fa28')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-9fa09613-00a3-4f41-b6a3-69361373fa28 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-9fa09613-00a3-4f41-b6a3-69361373fa28');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" Cpu_Core Cpu_Brand OpSys Storage_Type Inches Is_MacBook \\\n",
"0 -0.392649 -0.179219 -0.117135 -0.079704 0.068197 0.0865 \n",
"\n",
" Company Storage_GB Is_Touchscreen Weight_kg Type Gpu \\\n",
"0 0.149303 0.149772 0.191226 0.21037 0.292191 0.333036 \n",
"\n",
" Cpu_GHz ScreenResolution Screen_Height Screen_Width Ram \n",
"0 0.430293 0.453002 0.552809 0.556529 0.743007 "
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"data = Data(csv_file)\n",
"data.display()\n",
"data.plot_corr_with_Y(\"Price_euros\")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "Nda8_Nd5vp8y"
},
"source": [
"### Normalization \n",
"\n",
"Should not effect correlations."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 867
},
"id": "rCmjvwyJrVuf",
"outputId": "923c808e-aca6-46bf-b079-256d34cb2888"
},
"outputs": [
{
"data": {
"text/html": [
"\n",
" <div id=\"df-0e2d1c47-48d4-4a7e-86f1-a97a7f59ec54\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Company</th>\n",
" <th>Is_MacBook</th>\n",
" <th>Type</th>\n",
" <th>Inches</th>\n",
" <th>ScreenResolution</th>\n",
" <th>Ram</th>\n",
" <th>Gpu</th>\n",
" <th>OpSys</th>\n",
" <th>Weight_kg</th>\n",
" <th>Price_euros</th>\n",
" <th>Cpu_GHz</th>\n",
" <th>Cpu_Brand</th>\n",
" <th>Cpu_Core</th>\n",
" <th>Storage_GB</th>\n",
" <th>Storage_Type</th>\n",
" <th>Screen_Width</th>\n",
" <th>Screen_Height</th>\n",
" <th>Is_Touchscreen</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>-1.420695</td>\n",
" <td>8.220642</td>\n",
" <td>-1.194112</td>\n",
" <td>-1.204407</td>\n",
" <td>-2.674856</td>\n",
" <td>-0.075195</td>\n",
" <td>-1.489321</td>\n",
" <td>-3.409660</td>\n",
" <td>-1.005283</td>\n",
" <td>1339.69</td>\n",
" <td>0.002426</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-1.033943</td>\n",
" <td>-0.929659</td>\n",
" <td>1.345362</td>\n",
" <td>1.860586</td>\n",
" <td>-0.415713</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>-1.420695</td>\n",
" <td>-0.121645</td>\n",
" <td>-1.194112</td>\n",
" <td>-1.204407</td>\n",
" <td>-1.171259</td>\n",
" <td>-0.075195</td>\n",
" <td>-0.757853</td>\n",
" <td>-3.409660</td>\n",
" <td>-1.050381</td>\n",
" <td>898.94</td>\n",
" <td>-0.985431</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-1.033943</td>\n",
" <td>-0.111735</td>\n",
" <td>-0.919776</td>\n",
" <td>-0.600648</td>\n",
" <td>-0.415713</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>-1.057151</td>\n",
" <td>-0.121645</td>\n",
" <td>-0.424232</td>\n",
" <td>0.408772</td>\n",
" <td>0.332338</td>\n",
" <td>-0.075195</td>\n",
" <td>-0.757853</td>\n",
" <td>-1.729968</td>\n",
" <td>-0.268684</td>\n",
" <td>575.00</td>\n",
" <td>0.397569</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-0.757564</td>\n",
" <td>-0.929659</td>\n",
" <td>0.050997</td>\n",
" <td>0.032241</td>\n",
" <td>-0.415713</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>-1.420695</td>\n",
" <td>8.220642</td>\n",
" <td>-1.194112</td>\n",
" <td>0.268495</td>\n",
" <td>-2.674856</td>\n",
" <td>1.498767</td>\n",
" <td>-0.026385</td>\n",
" <td>-3.409660</td>\n",
" <td>-0.313782</td>\n",
" <td>2537.45</td>\n",
" <td>0.792712</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.324463</td>\n",
" <td>-0.204805</td>\n",
" <td>-0.929659</td>\n",
" <td>1.992544</td>\n",
" <td>2.563795</td>\n",
" <td>-0.415713</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>-1.420695</td>\n",
" <td>8.220642</td>\n",
" <td>-1.194112</td>\n",
" <td>-1.204407</td>\n",
" <td>-2.674856</td>\n",
" <td>-0.075195</td>\n",
" <td>-1.489321</td>\n",
" <td>-3.409660</td>\n",
" <td>-1.005283</td>\n",
" <td>1803.60</td>\n",
" <td>1.582997</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-0.757564</td>\n",
" <td>-0.929659</td>\n",
" <td>1.345362</td>\n",
" <td>1.860586</td>\n",
" <td>-0.415713</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-0e2d1c47-48d4-4a7e-86f1-a97a7f59ec54')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-0e2d1c47-48d4-4a7e-86f1-a97a7f59ec54 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-0e2d1c47-48d4-4a7e-86f1-a97a7f59ec54');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" Company Is_MacBook Type Inches ScreenResolution Ram \\\n",
"0 -1.420695 8.220642 -1.194112 -1.204407 -2.674856 -0.075195 \n",
"1 -1.420695 -0.121645 -1.194112 -1.204407 -1.171259 -0.075195 \n",
"2 -1.057151 -0.121645 -0.424232 0.408772 0.332338 -0.075195 \n",
"3 -1.420695 8.220642 -1.194112 0.268495 -2.674856 1.498767 \n",
"4 -1.420695 8.220642 -1.194112 -1.204407 -2.674856 -0.075195 \n",
"\n",
" Gpu OpSys Weight_kg Price_euros Cpu_GHz Cpu_Brand Cpu_Core \\\n",
"0 -1.489321 -3.409660 -1.005283 1339.69 0.002426 -0.223631 -0.745554 \n",
"1 -0.757853 -3.409660 -1.050381 898.94 -0.985431 -0.223631 -0.745554 \n",
"2 -0.757853 -1.729968 -0.268684 575.00 0.397569 -0.223631 -0.745554 \n",
"3 -0.026385 -3.409660 -0.313782 2537.45 0.792712 -0.223631 -0.324463 \n",
"4 -1.489321 -3.409660 -1.005283 1803.60 1.582997 -0.223631 -0.745554 \n",
"\n",
" Storage_GB Storage_Type Screen_Width Screen_Height Is_Touchscreen \n",
"0 -1.033943 -0.929659 1.345362 1.860586 -0.415713 \n",
"1 -1.033943 -0.111735 -0.919776 -0.600648 -0.415713 \n",
"2 -0.757564 -0.929659 0.050997 0.032241 -0.415713 \n",
"3 -0.204805 -0.929659 1.992544 2.563795 -0.415713 \n",
"4 -0.757564 -0.929659 1.345362 1.860586 -0.415713 "
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <div id=\"df-203ed8ec-57db-4e11-b57b-98481e1e6fe0\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>count</th>\n",
" <th>mean</th>\n",
" <th>std</th>\n",
" <th>min</th>\n",
" <th>25%</th>\n",
" <th>50%</th>\n",
" <th>75%</th>\n",
" <th>max</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>Company</th>\n",
" <td>1303.0</td>\n",
" <td>-2.181252e-17</td>\n",
" <td>1.000384</td>\n",
" <td>-1.420695</td>\n",
" <td>-0.693607</td>\n",
" <td>0.033481</td>\n",
" <td>0.397025</td>\n",
" <td>5.123096</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Is_MacBook</th>\n",
" <td>1303.0</td>\n",
" <td>1.090626e-17</td>\n",
" <td>1.000384</td>\n",
" <td>-0.121645</td>\n",
" <td>-0.121645</td>\n",
" <td>-0.121645</td>\n",
" <td>-0.121645</td>\n",
" <td>8.220642</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Type</th>\n",
" <td>1303.0</td>\n",
" <td>1.036095e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-1.194112</td>\n",
" <td>-0.424232</td>\n",
" <td>-0.424232</td>\n",
" <td>1.115529</td>\n",
" <td>2.655289</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Inches</th>\n",
" <td>1303.0</td>\n",
" <td>3.091924e-15</td>\n",
" <td>1.000384</td>\n",
" <td>-3.448829</td>\n",
" <td>-0.713439</td>\n",
" <td>0.408772</td>\n",
" <td>0.408772</td>\n",
" <td>2.372641</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ScreenResolution</th>\n",
" <td>1303.0</td>\n",
" <td>-1.690470e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-2.674856</td>\n",
" <td>-1.171259</td>\n",
" <td>0.332338</td>\n",
" <td>0.332338</td>\n",
" <td>3.339531</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Ram</th>\n",
" <td>1303.0</td>\n",
" <td>-5.453129e-17</td>\n",
" <td>1.000384</td>\n",
" <td>-1.255667</td>\n",
" <td>-0.862176</td>\n",
" <td>-0.075195</td>\n",
" <td>-0.075195</td>\n",
" <td>10.942543</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Gpu</th>\n",
" <td>1303.0</td>\n",
" <td>6.543755e-17</td>\n",
" <td>1.000384</td>\n",
" <td>-1.489321</td>\n",
" <td>-0.757853</td>\n",
" <td>-0.757853</td>\n",
" <td>0.705084</td>\n",
" <td>5.825363</td>\n",
" </tr>\n",
" <tr>\n",
" <th>OpSys</th>\n",
" <td>1303.0</td>\n",
" <td>-2.290314e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-3.409660</td>\n",
" <td>-0.050275</td>\n",
" <td>-0.050275</td>\n",
" <td>-0.050275</td>\n",
" <td>4.988804</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Weight_kg</th>\n",
" <td>1303.0</td>\n",
" <td>-1.008829e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-2.027503</td>\n",
" <td>-0.809859</td>\n",
" <td>0.001904</td>\n",
" <td>0.392752</td>\n",
" <td>4.000586</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Price_euros</th>\n",
" <td>1303.0</td>\n",
" <td>1.123687e+03</td>\n",
" <td>699.009043</td>\n",
" <td>174.000000</td>\n",
" <td>599.000000</td>\n",
" <td>977.000000</td>\n",
" <td>1487.880000</td>\n",
" <td>6099.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Cpu_GHz</th>\n",
" <td>1303.0</td>\n",
" <td>-2.562971e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-2.763574</td>\n",
" <td>-0.590288</td>\n",
" <td>0.397569</td>\n",
" <td>0.792712</td>\n",
" <td>2.570854</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Cpu_Brand</th>\n",
" <td>1303.0</td>\n",
" <td>-5.725786e-17</td>\n",
" <td>1.000384</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.223631</td>\n",
" <td>8.882358</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Cpu_Core</th>\n",
" <td>1303.0</td>\n",
" <td>-1.772267e-17</td>\n",
" <td>1.000384</td>\n",
" <td>-0.745554</td>\n",
" <td>-0.745554</td>\n",
" <td>-0.324463</td>\n",
" <td>0.517719</td>\n",
" <td>4.307540</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Storage_GB</th>\n",
" <td>1303.0</td>\n",
" <td>1.036095e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-1.293049</td>\n",
" <td>-0.757564</td>\n",
" <td>-0.230716</td>\n",
" <td>0.848891</td>\n",
" <td>4.113621</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Storage_Type</th>\n",
" <td>1303.0</td>\n",
" <td>4.089847e-17</td>\n",
" <td>1.000384</td>\n",
" <td>-0.929659</td>\n",
" <td>-0.929659</td>\n",
" <td>-0.111735</td>\n",
" <td>0.706189</td>\n",
" <td>3.977885</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Screen_Width</th>\n",
" <td>1303.0</td>\n",
" <td>1.417814e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-1.069437</td>\n",
" <td>-0.596185</td>\n",
" <td>0.050997</td>\n",
" <td>0.050997</td>\n",
" <td>3.934090</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Screen_Height</th>\n",
" <td>1303.0</td>\n",
" <td>-2.672033e-16</td>\n",
" <td>1.000384</td>\n",
" <td>-1.064766</td>\n",
" <td>-0.600648</td>\n",
" <td>0.032241</td>\n",
" <td>0.032241</td>\n",
" <td>3.829573</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Is_Touchscreen</th>\n",
" <td>1303.0</td>\n",
" <td>-2.726565e-18</td>\n",
" <td>1.000384</td>\n",
" <td>-0.415713</td>\n",
" <td>-0.415713</td>\n",
" <td>-0.415713</td>\n",
" <td>-0.415713</td>\n",
" <td>2.405506</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-203ed8ec-57db-4e11-b57b-98481e1e6fe0')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-203ed8ec-57db-4e11-b57b-98481e1e6fe0 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-203ed8ec-57db-4e11-b57b-98481e1e6fe0');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" count mean std min 25% \\\n",
"Company 1303.0 -2.181252e-17 1.000384 -1.420695 -0.693607 \n",
"Is_MacBook 1303.0 1.090626e-17 1.000384 -0.121645 -0.121645 \n",
"Type 1303.0 1.036095e-16 1.000384 -1.194112 -0.424232 \n",
"Inches 1303.0 3.091924e-15 1.000384 -3.448829 -0.713439 \n",
"ScreenResolution 1303.0 -1.690470e-16 1.000384 -2.674856 -1.171259 \n",
"Ram 1303.0 -5.453129e-17 1.000384 -1.255667 -0.862176 \n",
"Gpu 1303.0 6.543755e-17 1.000384 -1.489321 -0.757853 \n",
"OpSys 1303.0 -2.290314e-16 1.000384 -3.409660 -0.050275 \n",
"Weight_kg 1303.0 -1.008829e-16 1.000384 -2.027503 -0.809859 \n",
"Price_euros 1303.0 1.123687e+03 699.009043 174.000000 599.000000 \n",
"Cpu_GHz 1303.0 -2.562971e-16 1.000384 -2.763574 -0.590288 \n",
"Cpu_Brand 1303.0 -5.725786e-17 1.000384 -0.223631 -0.223631 \n",
"Cpu_Core 1303.0 -1.772267e-17 1.000384 -0.745554 -0.745554 \n",
"Storage_GB 1303.0 1.036095e-16 1.000384 -1.293049 -0.757564 \n",
"Storage_Type 1303.0 4.089847e-17 1.000384 -0.929659 -0.929659 \n",
"Screen_Width 1303.0 1.417814e-16 1.000384 -1.069437 -0.596185 \n",
"Screen_Height 1303.0 -2.672033e-16 1.000384 -1.064766 -0.600648 \n",
"Is_Touchscreen 1303.0 -2.726565e-18 1.000384 -0.415713 -0.415713 \n",
"\n",
" 50% 75% max \n",
"Company 0.033481 0.397025 5.123096 \n",
"Is_MacBook -0.121645 -0.121645 8.220642 \n",
"Type -0.424232 1.115529 2.655289 \n",
"Inches 0.408772 0.408772 2.372641 \n",
"ScreenResolution 0.332338 0.332338 3.339531 \n",
"Ram -0.075195 -0.075195 10.942543 \n",
"Gpu -0.757853 0.705084 5.825363 \n",
"OpSys -0.050275 -0.050275 4.988804 \n",
"Weight_kg 0.001904 0.392752 4.000586 \n",
"Price_euros 977.000000 1487.880000 6099.000000 \n",
"Cpu_GHz 0.397569 0.792712 2.570854 \n",
"Cpu_Brand -0.223631 -0.223631 8.882358 \n",
"Cpu_Core -0.324463 0.517719 4.307540 \n",
"Storage_GB -0.230716 0.848891 4.113621 \n",
"Storage_Type -0.111735 0.706189 3.977885 \n",
"Screen_Width 0.050997 0.050997 3.934090 \n",
"Screen_Height 0.032241 0.032241 3.829573 \n",
"Is_Touchscreen -0.415713 -0.415713 2.405506 "
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"data.norm(\"Price_euros\")\n",
"data.display()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 558
},
"id": "EUWVDgvTw44p",
"outputId": "83880271-f5f8-4f6d-d24b-f6cb63f9fbd3"
},
"outputs": [
{
"data": {
"text/html": [
"\n",
" <div id=\"df-a63db2cc-2908-4efc-bcc8-e62543775987\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Cpu_Core</th>\n",
" <th>Cpu_Brand</th>\n",
" <th>OpSys</th>\n",
" <th>Storage_Type</th>\n",
" <th>Inches</th>\n",
" <th>Is_MacBook</th>\n",
" <th>Company</th>\n",
" <th>Storage_GB</th>\n",
" <th>Is_Touchscreen</th>\n",
" <th>Weight_kg</th>\n",
" <th>Type</th>\n",
" <th>Gpu</th>\n",
" <th>Cpu_GHz</th>\n",
" <th>ScreenResolution</th>\n",
" <th>Screen_Height</th>\n",
" <th>Screen_Width</th>\n",
" <th>Ram</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>-0.392649</td>\n",
" <td>-0.179219</td>\n",
" <td>-0.117135</td>\n",
" <td>-0.079704</td>\n",
" <td>0.068197</td>\n",
" <td>0.0865</td>\n",
" <td>0.149303</td>\n",
" <td>0.149772</td>\n",
" <td>0.191226</td>\n",
" <td>0.21037</td>\n",
" <td>0.292191</td>\n",
" <td>0.333036</td>\n",
" <td>0.430293</td>\n",
" <td>0.453002</td>\n",
" <td>0.552809</td>\n",
" <td>0.556529</td>\n",
" <td>0.743007</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a63db2cc-2908-4efc-bcc8-e62543775987')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-a63db2cc-2908-4efc-bcc8-e62543775987 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-a63db2cc-2908-4efc-bcc8-e62543775987');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" Cpu_Core Cpu_Brand OpSys Storage_Type Inches Is_MacBook \\\n",
"0 -0.392649 -0.179219 -0.117135 -0.079704 0.068197 0.0865 \n",
"\n",
" Company Storage_GB Is_Touchscreen Weight_kg Type Gpu \\\n",
"0 0.149303 0.149772 0.191226 0.21037 0.292191 0.333036 \n",
"\n",
" Cpu_GHz ScreenResolution Screen_Height Screen_Width Ram \n",
"0 0.430293 0.453002 0.552809 0.556529 0.743007 "
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"data.plot_corr_with_Y(\"Price_euros\")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "lxRX4VL0wGH5"
},
"source": [
"## Regression "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "dgz-T1ctpoUz"
},
"outputs": [],
"source": [
"class ModelTrainer:\n",
" def __init__(self, data, Y_name, test_size=.3, random_state=42, show_samples=False, \n",
" grid_search_cv=3, grid_search_scoring=None, grid_search_refit=None): \n",
" data_copy = data.copy()\n",
" Y = data_copy[Y_name]\n",
" X = data_copy.drop(Y_name, 1)\n",
"\n",
" self.n_features = len(X.columns)\n",
" self.random_state = random_state\n",
" self.grid_search_cv=grid_search_cv\n",
" self.grid_search_refit=grid_search_refit if grid_search_refit!=None else \"neg_root_mean_squared_error\"\n",
" self.grid_search_scoring=grid_search_scoring if grid_search_scoring!=None else [\"neg_root_mean_squared_error\", \"r2\"]\n",
"\n",
" self.X_train, self.X_test, self.y_train, self.y_test = train_test_split(\n",
" X.values, Y.values, \n",
" test_size=test_size, random_state=random_state)\n",
" \n",
" self.models={}\n",
" self.metrics={\"RMSE\":{}, \"R2_Score\":{}}\n",
" \n",
" if show_samples:\n",
" self.sampling()\n",
"\n",
" def sampling(self, rows=5):\n",
" display(self.X_train[:rows])\n",
" display(self.X_test[:rows])\n",
" display(self.y_train[:rows])\n",
" display(self.y_test[:rows])\n",
"\n",
" def rmse(self, yTrue, yPreds):\n",
" return np.sqrt(mean_squared_error(yTrue, yPreds))\n",
"\n",
" def fit(self, model, name):\n",
" model = model.fit(self.X_train, self.y_train)\n",
"\n",
" yPreds = model.best_estimator_.predict(self.X_test) if isinstance(model, GridSearchCV) else model.predict(self.X_test)\n",
"\n",
" self.models[name] = model\n",
" self.metrics[\"RMSE\"][name] = self.rmse(self.y_test, yPreds)\n",
" self.metrics[\"R2_Score\"][name] = r2_score(self.y_test, yPreds)\n",
"\n",
" def _create_GridSearchCV(self, estimator, param_grid): \n",
" return GridSearchCV(\n",
" estimator,\n",
" param_grid=param_grid,\n",
" scoring=self.grid_search_scoring,\n",
" cv=self.grid_search_cv,\n",
" refit=self.grid_search_refit,\n",
" n_jobs=-1)\n",
"\n",
" def linear_regression(self):\n",
" self.fit(LinearRegression(), \"linear_regression\")\n",
"\n",
" def ridge(self):\n",
" param_grid = {\n",
" 'alpha': np.geomspace(1e-9, 1e0, num=10),\n",
" 'max_iter': [1e3, 1e4, 1e5, 1e6]\n",
" }\n",
" \n",
" self.fit(self._create_GridSearchCV(\n",
" Ridge(),\n",
" param_grid=param_grid), \"ridge\")\n",
"\n",
" def lasso(self):\n",
" param_grid = {\n",
" 'alpha': np.geomspace(1e-9, 1e0, num=10),\n",
" 'max_iter': [1e3, 1e4, 1e5, 1e6]\n",
" }\n",
" \n",
" self.fit(self._create_GridSearchCV(\n",
" Lasso(),\n",
" param_grid=param_grid), \"lasso\")\n",
" \n",
" def elastic_net(self):\n",
" param_grid = {\n",
" 'alpha': np.geomspace(1e-9, 1e0, num=10),\n",
" 'l1_ratio': np.linspace(0.1, 0.9, 9),\n",
" 'max_iter': [1e3, 1e4, 1e5, 1e6]\n",
" }\n",
" \n",
" self.fit(self._create_GridSearchCV(\n",
" ElasticNet(),\n",
" param_grid=param_grid), \"elastic_net\")\n",
" \n",
" def knn(self, max_k=50):\n",
" param_grid = {\n",
" 'n_neighbors': list(range(1, max_k+1)),\n",
" 'weights': ['uniform', 'distance'],\n",
" }\n",
"\n",
" self.fit(self._create_GridSearchCV(\n",
" KNeighborsRegressor(),\n",
" param_grid=param_grid), \"KNN\")\n",
"\n",
" def decision_tree(self, use_grid_search=True, list_params=False):\n",
" model = DecisionTreeRegressor()\n",
" if list_params:\n",
" display(pd.Series(model.get_params().keys(), name=\"decision_tree params\").to_frame())\n",
"\n",
" self.fit(model, \"decision_tree\")\n",
"\n",
" # The grid model will replace previous model.\n",
" if use_grid_search:\n",
" param_grid = {\n",
" 'max_depth':range(1, model.tree_.max_depth+1, 2),\n",
" 'max_features': range(1, len(model.feature_importances_)+1), \n",
" } \n",
" \n",
" self.fit(self._create_GridSearchCV(\n",
" DecisionTreeRegressor(random_state=self.random_state),\n",
" param_grid=param_grid), \"decision_tree\")\n",
" \n",
" def _get_tree_base_estimator(self, use_fit_tree):\n",
" if use_fit_tree and self.models.get(\"decision_tree\") != None:\n",
" return self.models[\"decision_tree\"]\n",
" else:\n",
" return DecisionTreeRegressor(random_state=self.random_state) \n",
"\n",
" def _ensemble_fit(self, estimator, use_grid_search, param_grid, name):\n",
" if use_grid_search: \n",
" self.fit(self._create_GridSearchCV(\n",
" estimator,\n",
" param_grid=param_grid), name)\n",
" else:\n",
" self.fit(estimator, name) \n",
"\n",
" def random_forest(self, list_params=False):\n",
" rf_estimator = RandomForestRegressor() \n",
" if list_params:\n",
" display(pd.Series(rf_estimator.get_params().keys(), \n",
" name=\"random_forest params\").to_frame())\n",
" \n",
" param_grid = {'n_estimators': [2*n+1 for n in range(20)],\n",
" 'max_depth': [n for n in range(1, self.n_features+1)],\n",
" 'max_features': [n for n in range(1, self.n_features+1)],\n",
" } \n",
" self._ensemble_fit(rf_estimator, True, param_grid, \"random_forest\")\n",
"\n",
" def bagging(self, use_fit_tree=True, use_grid_search=True, list_params=False):\n",
" base_estimator = self._get_tree_base_estimator(use_fit_tree) \n",
" if list_params:\n",
" display(pd.Series(base_estimator.get_params().keys(), \n",
" name=\"bagging params\").to_frame())\n",
" \n",
" param_grid = {'n_estimators': [2*n+1 for n in range(20)]} if use_fit_tree \\\n",
" else {'n_estimators': [2*n+1 for n in range(20)],\n",
" 'base_estimator__max_features': [n for n in range(1, self.n_features+1)],\n",
" 'base_estimator__max_depth' : [2*n+1 for n in range(1, 10) ] \n",
" }\n",
" bagging_estimator = BaggingRegressor(base_estimator=base_estimator, \n",
" random_state=self.random_state,\n",
" bootstrap=True)\n",
" self._ensemble_fit(bagging_estimator, use_grid_search, param_grid, \"bagging\")\n",
"\n",
" def ada_boost(self, use_fit_tree=True, use_grid_search=True, list_params=False):\n",
" base_estimator = self._get_tree_base_estimator(use_fit_tree) \n",
" if list_params:\n",
" display(pd.Series(base_estimator.get_params().keys(), \n",
" name=\"ada_boost params\").to_frame())\n",
" \n",
" param_grid = {'n_estimators': [2*n+1 for n in range(20)]} if use_fit_tree \\\n",
" else {'n_estimators': [2*n+1 for n in range(20)],\n",
" 'learning_rate': [0.1*(n+1) for n in range(10)],#[1e-3],#[0.1*(n+1) for n in range(10)],\n",
" 'base_estimator__max_features': [n for n in range(1, self.n_features+1)],\n",
" 'base_estimator__max_depth' : [2*n+1 for n in range(1, 10) ] \n",
" }\n",
" ada_boost_estimator = AdaBoostRegressor(base_estimator=base_estimator,\n",
" random_state=self.random_state)\n",
" self._ensemble_fit(ada_boost_estimator, use_grid_search, param_grid, \"ada_boost\")\n",
"\n",
" def gradient_boosting(self, list_params=False):\n",
" GB_estimator = GradientBoostingRegressor() \n",
" if list_params:\n",
" display(pd.Series(GB_estimator.get_params().keys(), \n",
" name=\"gradient_boosting params\").to_frame())\n",
" \n",
" param_grid = {'n_estimators': [2*n+1 for n in range(20)],\n",
" 'learning_rate': [0.1*(n+1) for n in range(10)],#[1e-3],#[0.1*(n+1) for n in range(10)],\n",
" 'subsample': [1.0, 0.5], #[1.0],# [1.0, 0.5],\n",
" 'max_features': [n for n in range(1, self.n_features+1)],\n",
" 'max_depth': [2*n+1 for n in range(1, 10) ] \n",
" }\n",
" \n",
" self._ensemble_fit(GB_estimator, True, param_grid, \"gradient_boosting\")\n",
"\n",
" def voting(self, verbose=True):\n",
" estimators = [(k,v) for k,v in self.models.items()]\n",
" self.fit(VotingRegressor(estimators, verbose=verbose, n_jobs=-1), \"Voting\")\n",
"\n",
" def show_metrics(self, metrics_table): metrics_table(self.metrics)\n",
"\n",
" def plot_prediction(self, plot_preds): plot_preds(self.models, self.X_test, self. y_test)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "zXSkarglsnyf"
},
"outputs": [],
"source": [
"class PredictionPlot:\n",
" def __init__(self, figsize=(5, 5)) -> None:\n",
" self.figsize = figsize\n",
"\n",
" def __call__(self, \n",
" models, \n",
" X, y):\n",
" fig = plt.figure(figsize=self.figsize)\n",
" ax = plt.axes()\n",
"\n",
" for label, model in models.items():\n",
" ax.plot(y, \n",
" model.predict(X), \n",
" marker='o', ls='', ms=3.0, \n",
" label=label)\n",
"\n",
" lim = (0, y.max())\n",
" ax.set(xlabel='Actual Price', \n",
" ylabel='Predicted Price', \n",
" xlim=lim,\n",
" ylim=lim,\n",
" title='Regression Results');\n",
"\n",
" leg = plt.legend(frameon=True)\n",
" leg.get_frame().set_edgecolor('black')\n",
" leg.get_frame().set_linewidth(1.0)\n",
" plt.tight_layout()\n",
" plt.show()\n",
"\n",
"\n",
" for label, model in models.items():\n",
" fig = plt.figure(figsize=self.figsize)\n",
" ax = plt.axes()\n",
" ax.plot(y, \n",
" model.predict(X), \n",
" marker='o', ls='', ms=3.0, \n",
" label=label)\n",
" leg = plt.legend(frameon=True)\n",
" leg.get_frame().set_edgecolor('black')\n",
" leg.get_frame().set_linewidth(1.0)\n",
" plt.tight_layout()\n",
" plt.show()\n",
"\n",
"plot_preds = PredictionPlot()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "7lRZbU6szppm"
},
"outputs": [],
"source": [
"class MetricsTable:\n",
" def __init__(self, figsize=(15, 5)) -> None:\n",
" self.figsize = figsize\n",
"\n",
" def __call__(self, metrics):\n",
" df = pd.DataFrame([metrics[\"RMSE\"], \n",
" metrics[\"R2_Score\"]], \n",
" index=[\"RMSE\", \"R2_Score\"]).T\n",
" display(df)\n",
"\n",
" colors_ = [c for c, _ in mcolors.CSS4_COLORS.items()]\n",
" color_list=list(map(lambda idx: colors_[(idx+10)%len(colors_)], pd.factorize(df.index.unique())[0]))\n",
" \n",
" fig, axs=plt.subplots(nrows=2, ncols=1, figsize=self.figsize)\n",
"\n",
" df[\"RMSE\"].sort_values().plot.bar(color=color_list, rot=360, ax=axs[0], width=.25)\n",
" axs[0].set_title(f\"RMSE\")\n",
" \n",
" df[\"R2_Score\"].sort_values().plot.bar(color=color_list, rot=360, ax=axs[1], width=.25)\n",
" axs[1].set_title(f\"R2_Score\")\n",
"\n",
" plt.tight_layout()\n",
" plt.show()\n",
"\n",
"metrics_table = MetricsTable()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "33Uz3hbsaXjj"
},
"source": [
"## Feature selection via Lasso (Optional)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
},
"id": "PZXLxKxsanoW",
"outputId": "31422a14-81a3-4463-c1a1-2f901f767ba7"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:6: FutureWarning: In a future version of pandas all arguments of DataFrame.drop except for the argument 'labels' will be keyword-only\n",
" \n"
]
},
{
"data": {
"text/html": [
"\n",
" <div id=\"df-583c3aa2-0954-460b-a0f6-ba006ce18371\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>0</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>15.514539</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>-3.178560</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>53.483882</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>-86.016652</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>-100.428163</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>379.877332</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>74.337158</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>-10.938261</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>40.800886</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>127.384061</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>-63.956323</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>-42.816655</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td>-20.374371</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13</th>\n",
" <td>-65.146773</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14</th>\n",
" <td>224.654579</td>\n",
" </tr>\n",
" <tr>\n",
" <th>15</th>\n",
" <td>-0.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>16</th>\n",
" <td>-0.000000</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-583c3aa2-0954-460b-a0f6-ba006ce18371')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-583c3aa2-0954-460b-a0f6-ba006ce18371 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-583c3aa2-0954-460b-a0f6-ba006ce18371');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" 0\n",
"0 15.514539\n",
"1 -3.178560\n",
"2 53.483882\n",
"3 -86.016652\n",
"4 -100.428163\n",
"5 379.877332\n",
"6 74.337158\n",
"7 -10.938261\n",
"8 40.800886\n",
"9 127.384061\n",
"10 -63.956323\n",
"11 -42.816655\n",
"12 -20.374371\n",
"13 -65.146773\n",
"14 224.654579\n",
"15 -0.000000\n",
"16 -0.000000"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:11: FutureWarning: In a future version of pandas all arguments of DataFrame.drop except for the argument 'labels' will be keyword-only\n",
" # This is added back by InteractiveShellApp.init_path()\n"
]
},
{
"data": {
"text/html": [
"\n",
" <div id=\"df-e48fee83-a330-4e1d-8ac5-939754345f54\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>0</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>Company</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>Is_MacBook</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>Type</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>Inches</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>ScreenResolution</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>Ram</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>Gpu</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>OpSys</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>Weight_kg</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>Cpu_GHz</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>Cpu_Brand</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>Cpu_Core</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td>Storage_GB</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13</th>\n",
" <td>Storage_Type</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14</th>\n",
" <td>Screen_Width</td>\n",
" </tr>\n",
" <tr>\n",
" <th>15</th>\n",
" <td>Screen_Height</td>\n",
" </tr>\n",
" <tr>\n",
" <th>16</th>\n",
" <td>Is_Touchscreen</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-e48fee83-a330-4e1d-8ac5-939754345f54')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-e48fee83-a330-4e1d-8ac5-939754345f54 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-e48fee83-a330-4e1d-8ac5-939754345f54');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" 0\n",
"0 Company\n",
"1 Is_MacBook\n",
"2 Type\n",
"3 Inches\n",
"4 ScreenResolution\n",
"5 Ram\n",
"6 Gpu\n",
"7 OpSys\n",
"8 Weight_kg\n",
"9 Cpu_GHz\n",
"10 Cpu_Brand\n",
"11 Cpu_Core\n",
"12 Storage_GB\n",
"13 Storage_Type\n",
"14 Screen_Width\n",
"15 Screen_Height\n",
"16 Is_Touchscreen"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"model_trainer = ModelTrainer(data.data, \"Price_euros\")\n",
"\n",
"model_trainer.lasso()\n",
"max_iter = model_trainer.models[\"lasso\"].estimator.max_iter\n",
"alpha = model_trainer.models[\"lasso\"].estimator.alpha\n",
"model_trainer.fit(Lasso(alpha=alpha, max_iter=max_iter), \"feature selector\")\n",
"coef = model_trainer.models[\"feature selector\"].coef_\n",
"display(pd.Series(coef).to_frame())\n",
"\n",
"data_copy = data.data.copy()\n",
"X = data_copy.drop(\"Price_euros\", 1)\n",
"display(pd.Series(X.columns).to_frame())"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "PKOsH8n2egXS"
},
"source": [
"### Feature cut\n",
"\n",
"It turns out that features like `Is_MacBook`, `Screen_Height` and `Is_Touchscreen` are quite \n",
"unimportant to the price as target (cofe_<3), we could cut them and fit the models."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 206
},
"id": "TuA7RtWQe_-K",
"outputId": "5864203e-63c9-474b-9ab3-3e04fda58c77"
},
"outputs": [
{
"data": {
"text/html": [
"\n",
" <div id=\"df-b949568f-86d2-40fa-b00c-60703be5186d\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Company</th>\n",
" <th>Type</th>\n",
" <th>Inches</th>\n",
" <th>ScreenResolution</th>\n",
" <th>Ram</th>\n",
" <th>Gpu</th>\n",
" <th>OpSys</th>\n",
" <th>Weight_kg</th>\n",
" <th>Price_euros</th>\n",
" <th>Cpu_GHz</th>\n",
" <th>Cpu_Brand</th>\n",
" <th>Cpu_Core</th>\n",
" <th>Storage_GB</th>\n",
" <th>Storage_Type</th>\n",
" <th>Screen_Width</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>-1.420695</td>\n",
" <td>-1.194112</td>\n",
" <td>-1.204407</td>\n",
" <td>-2.674856</td>\n",
" <td>-0.075195</td>\n",
" <td>-1.489321</td>\n",
" <td>-3.409660</td>\n",
" <td>-1.005283</td>\n",
" <td>1339.69</td>\n",
" <td>0.002426</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-1.033943</td>\n",
" <td>-0.929659</td>\n",
" <td>1.345362</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>-1.420695</td>\n",
" <td>-1.194112</td>\n",
" <td>-1.204407</td>\n",
" <td>-1.171259</td>\n",
" <td>-0.075195</td>\n",
" <td>-0.757853</td>\n",
" <td>-3.409660</td>\n",
" <td>-1.050381</td>\n",
" <td>898.94</td>\n",
" <td>-0.985431</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-1.033943</td>\n",
" <td>-0.111735</td>\n",
" <td>-0.919776</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>-1.057151</td>\n",
" <td>-0.424232</td>\n",
" <td>0.408772</td>\n",
" <td>0.332338</td>\n",
" <td>-0.075195</td>\n",
" <td>-0.757853</td>\n",
" <td>-1.729968</td>\n",
" <td>-0.268684</td>\n",
" <td>575.00</td>\n",
" <td>0.397569</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-0.757564</td>\n",
" <td>-0.929659</td>\n",
" <td>0.050997</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>-1.420695</td>\n",
" <td>-1.194112</td>\n",
" <td>0.268495</td>\n",
" <td>-2.674856</td>\n",
" <td>1.498767</td>\n",
" <td>-0.026385</td>\n",
" <td>-3.409660</td>\n",
" <td>-0.313782</td>\n",
" <td>2537.45</td>\n",
" <td>0.792712</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.324463</td>\n",
" <td>-0.204805</td>\n",
" <td>-0.929659</td>\n",
" <td>1.992544</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>-1.420695</td>\n",
" <td>-1.194112</td>\n",
" <td>-1.204407</td>\n",
" <td>-2.674856</td>\n",
" <td>-0.075195</td>\n",
" <td>-1.489321</td>\n",
" <td>-3.409660</td>\n",
" <td>-1.005283</td>\n",
" <td>1803.60</td>\n",
" <td>1.582997</td>\n",
" <td>-0.223631</td>\n",
" <td>-0.745554</td>\n",
" <td>-0.757564</td>\n",
" <td>-0.929659</td>\n",
" <td>1.345362</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b949568f-86d2-40fa-b00c-60703be5186d')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-b949568f-86d2-40fa-b00c-60703be5186d button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-b949568f-86d2-40fa-b00c-60703be5186d');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" Company Type Inches ScreenResolution Ram Gpu \\\n",
"0 -1.420695 -1.194112 -1.204407 -2.674856 -0.075195 -1.489321 \n",
"1 -1.420695 -1.194112 -1.204407 -1.171259 -0.075195 -0.757853 \n",
"2 -1.057151 -0.424232 0.408772 0.332338 -0.075195 -0.757853 \n",
"3 -1.420695 -1.194112 0.268495 -2.674856 1.498767 -0.026385 \n",
"4 -1.420695 -1.194112 -1.204407 -2.674856 -0.075195 -1.489321 \n",
"\n",
" OpSys Weight_kg Price_euros Cpu_GHz Cpu_Brand Cpu_Core \\\n",
"0 -3.409660 -1.005283 1339.69 0.002426 -0.223631 -0.745554 \n",
"1 -3.409660 -1.050381 898.94 -0.985431 -0.223631 -0.745554 \n",
"2 -1.729968 -0.268684 575.00 0.397569 -0.223631 -0.745554 \n",
"3 -3.409660 -0.313782 2537.45 0.792712 -0.223631 -0.324463 \n",
"4 -3.409660 -1.005283 1803.60 1.582997 -0.223631 -0.745554 \n",
"\n",
" Storage_GB Storage_Type Screen_Width \n",
"0 -1.033943 -0.929659 1.345362 \n",
"1 -1.033943 -0.111735 -0.919776 \n",
"2 -0.757564 -0.929659 0.050997 \n",
"3 -0.204805 -0.929659 1.992544 \n",
"4 -0.757564 -0.929659 1.345362 "
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"data_feature_cut = data_copy.drop([\"Is_MacBook\", \"Screen_Height\", \"Is_Touchscreen\"], axis = 1)\n",
"display(data_feature_cut.head())"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "9ckefK_Yf_sA"
},
"source": [
"## Fit models"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "ufqvxaE4gD6i",
"outputId": "0d5ec89f-57be-4d4d-daff-5065d98975f7"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:6: FutureWarning: In a future version of pandas all arguments of DataFrame.drop except for the argument 'labels' will be keyword-only\n",
" \n"
]
}
],
"source": [
"model_trainer = ModelTrainer(data_feature_cut, \"Price_euros\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "ysULCdaRfyji"
},
"outputs": [],
"source": [
"model_trainer.knn()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "9OyBBo05yNDT"
},
"outputs": [],
"source": [
"model_trainer.linear_regression()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "4ymlme0oZkHg"
},
"outputs": [],
"source": [
"model_trainer.lasso()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "abUZoS5aZj3W"
},
"outputs": [],
"source": [
"model_trainer.ridge()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "ddP1Rx8eZjoS"
},
"outputs": [],
"source": [
"model_trainer.elastic_net()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "BdxYroBCBqps"
},
"outputs": [],
"source": [
"# list_params=True, we can see the params which can be passed to GridSearchCV.\n",
"model_trainer.decision_tree()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "zqCB20rWitZ9"
},
"outputs": [],
"source": [
"# list_params=True, we can see the params which can be passed to GridSearchCV.\n",
"model_trainer.random_forest()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "yK9iOk5eBqf8"
},
"outputs": [],
"source": [
"# list_params=True, we can see the params which can be passed to GridSearchCV.\n",
"model_trainer.bagging(use_fit_tree=False) "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "vADKHjJ4QBwm"
},
"outputs": [],
"source": [
"# list_params=True, we can see the params which can be passed to GridSearchCV.\n",
"model_trainer.ada_boost(use_fit_tree=False) "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "mW2D3YkcpSW7"
},
"outputs": [],
"source": [
"# list_params=True, we can see the params which can be passed to GridSearchCV.\n",
"model_trainer.gradient_boosting() "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "4tR9ikJY3sS4"
},
"outputs": [],
"source": [
"model_trainer.voting() "
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "W3d_ehTKKHDY"
},
"source": [
"## Prediction on unseen test data"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 746
},
"id": "E7m-Fbj_9DhS",
"outputId": "9333ed29-dd90-4b24-bbff-e6d0101b8def"
},
"outputs": [
{
"data": {
"text/html": [
"\n",
" <div id=\"df-6785692c-8edd-443b-bf13-7f3dfda12623\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>RMSE</th>\n",
" <th>R2_Score</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>KNN</th>\n",
" <td>340.583723</td>\n",
" <td>0.783377</td>\n",
" </tr>\n",
" <tr>\n",
" <th>linear_regression</th>\n",
" <td>414.135235</td>\n",
" <td>0.679712</td>\n",
" </tr>\n",
" <tr>\n",
" <th>lasso</th>\n",
" <td>414.155378</td>\n",
" <td>0.679680</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ridge</th>\n",
" <td>414.113387</td>\n",
" <td>0.679745</td>\n",
" </tr>\n",
" <tr>\n",
" <th>elastic_net</th>\n",
" <td>413.999128</td>\n",
" <td>0.679922</td>\n",
" </tr>\n",
" <tr>\n",
" <th>decision_tree</th>\n",
" <td>406.948385</td>\n",
" <td>0.690732</td>\n",
" </tr>\n",
" <tr>\n",
" <th>random_forest</th>\n",
" <td>307.363006</td>\n",
" <td>0.823575</td>\n",
" </tr>\n",
" <tr>\n",
" <th>bagging</th>\n",
" <td>305.001564</td>\n",
" <td>0.826276</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ada_boost</th>\n",
" <td>311.341650</td>\n",
" <td>0.818978</td>\n",
" </tr>\n",
" <tr>\n",
" <th>gradient_boosting</th>\n",
" <td>298.475108</td>\n",
" <td>0.833631</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Voting</th>\n",
" <td>313.899631</td>\n",
" <td>0.815991</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6785692c-8edd-443b-bf13-7f3dfda12623')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-6785692c-8edd-443b-bf13-7f3dfda12623 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-6785692c-8edd-443b-bf13-7f3dfda12623');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" RMSE R2_Score\n",
"KNN 340.583723 0.783377\n",
"linear_regression 414.135235 0.679712\n",
"lasso 414.155378 0.679680\n",
"ridge 414.113387 0.679745\n",
"elastic_net 413.999128 0.679922\n",
"decision_tree 406.948385 0.690732\n",
"random_forest 307.363006 0.823575\n",
"bagging 305.001564 0.826276\n",
"ada_boost 311.341650 0.818978\n",
"gradient_boosting 298.475108 0.833631\n",
"Voting 313.899631 0.815991"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 1080x360 with 2 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"model_trainer.show_metrics(metrics_table)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
},
"id": "C7wLO-D39ScI",
"outputId": "21263c8c-b05e-4049-cbd5-d136b8219521"
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAFgCAYAAACFYaNMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU1fn48c9zZyaBJGSBEMISSCIEyqIsYRGLWhdQVFAUtaJFUVyr1UqtShe//aJtLdWfu18VBa0WKS5FcQVBUWRJ2EWWGBLDTiAJSUhmuff8/rg3YwJJyD4Ezvv1ymvm3rnn3jMTeHLm3HOeI0opNE3TtJZnhLoCmqZppyodgDVN00JEB2BN07QQ0QFY0zQtRHQA1jRNCxEdgDVN00JEB2DtlCEio0Rka6jr0VRERIlIz1DXQ2s4HYC1ehGRHBEpE5ESEdkrIrNFJCrU9aoLpdQypVTvpj6viCQ7wbDE+ckRkQeb+jrHqcNsEZnRktfUGk8HYK0hLlNKRQEDgUHAQ019ARFxN/U5W0Cs87lcBfxRRC4MdYW0E5sOwFqDKaX2Ap9iB2IARGSEiCwXkUIRWS8i51Z6LUVEvhKRYhFZJCLPici/nNcqWpE3i8iPwBfO/iki8r2IFIjIpyLSw9kvIvKkiOwXkcMislFE+juvjRWRzc51donINGf/uSKys1J9fiYiS526fici4yq9Ntup30LnPCtF5LQ6fi4ZwHdHfS4NeR9LReSWSue4UUS+Pvp6InIrMAl4wGmBf+Ds/73z/otFZKuInF+X+mstRwdgrcFEpBtwMZDlbHcFFgIzgPbANOAdEenoFHkLWAV0AB4BbqjmtOcAPwPGiMh44GFgAtARWAb82zluNHA2kAbEAFcDB53XZgG3KaXaAf1xgvlRdfcAHwCfAQnA3cCbIlK5i+Ja4H+AOOc9PlrHz2WEc92Kz6Wh76NOlFIvAW8CjyulopRSlznv49fAUOdzGAPk1Oe8WvPTAVhriPdFpBjIA/YDf3b2Xw98pJT6SCllKaU+BzKAsSLSHRgK/Ekp5VNKfQ0sqObcjyilSpVSZcDtwF+VUt8rpQLAY8BAp/XoB9oBfQBxjtnjnMMP9BWRaKVUgVJqTTXXGQFEAX9z6vMF8CHwy0rHvKeUWuVc+00qtWhrkC8iZcC3wPPA+87+hr6PxjCBcOzPwaOUylFK/dAE59WakA7AWkNc7rSqzsUOHPHO/h7AROcrfaGIFAI/BzoDXYBDSqkjlc6TV825K+/rATxV6VyHAAG6OgHzWeA5YL+IvCQi0U65K4GxQK6IfCkiZ1ZznS5AnlLKqrQvF+haaXtvpedHsAN2beKdY+7H/mw8jXwfDaaUygLuxf6msV9E5opIl8aeV2taOgBrDaaU+hKYDcx0duUBbyilYiv9RCql/gbsAdqLSESlUyRVd9pKz/OwuxIqn6+tUmq5c/2nlVJDgL7YX+F/5+xfrZQaj9218D4wr5rr7AaSRKTy/4HuwK56fQhHV14pUyn1BFAO3NmY9wGUApU/r8TaLl1NXd5SSv0c+w+AAv7emPemNT0dgLXG+n/AhSJyBvAv4DIRGSMiLhFp49z46qaUysXujnhERMKcVullxzn3i8BDItIPQERiRGSi83yoiAx3+nJLsQOe5Zx7kojEKKX8wGHAqubcK7FbtQ+IiMe5WXgZMLeRn0eFvznnbtOQ9+GcYx0wQUQixB7ve3Mt19sHpFZsiEhvETlPRMKdc5ZR/eeghZAOwFqjKKUOAK9j9+3mARU3nA5gt/x+x0//ziYBZ2LfZJoBvA14azn3e9ittrkichjYhH3TDyAaeBkowO46OAj8w3ntBiDHKXO7c92jz+3DDrgXA/nYfba/UkptqfeHUL2FTt2mNuJ9PAn4sIPrHOx+6JrMwu7vLRSR97H7f//mvLe92N8Gmny4oNY4ohOya6EiIm8DW5RSfz7uwZp2EtItYK3FOF+3TxMRQ0Quwm4tv3+8cpp2smqNs4201isReBd7HPBO4A6l1NrQVknTQkd3QWiapoWI7oLQNE0LkZOyCyI+Pl4lJyeHuhqapp0iMjMz85VSHY9/ZFUnZQBOTk4mIyMj1NXQNO0UISK5DSnXrF0QIhIrIvNFZIuTCepMEWkvIp+LyHbnMc45VkTkaRHJEpENIjK40nkmO8dvF5HJzVlnTdO0ltLcfcBPAZ8opfoAZwDfAw8Ci5VSvYDFzjbYA9N7OT+3Ai8AiEh77GQvw4FhwJ8rgramaVpr1mwBWERisNPszQJ75pFSqhB77Occ57A5wOXO8/HA68q2AogVkc7YafQ+V0odUkoVAJ8DFzVXvTVN01pKc7aAU7Cno74mImtF5BURiQQ6VUq3txfo5DzvStVMWDudfTXtr0JEbhWRDBHJOHDgQBO/FU3TtKbXnAHYDQwGXlBKDcJONFJlnSxlD0JukoHISqmXlFLpSqn0jh3rfTNS0zStxTVnAN4J7FRKrXS252MH5H1O1wLO437n9V1UTU/YzdlX035N07RWrdkCsLNeWF6lJV7OBzZjr4JQMZJhMvBf5/kC4FfOaIgRQJHTVfEpMFpE4pybb6OdfZqmaa1ac48DrlhnKwzIBm7CDvrzRORm7PR7VzvHfoS9ikEWdp7WmwCUUodE5H+B1c5xf1FKHWrmemuapjW7kzIXRHp6utITMTRNaykikqmUSq9vOZ0LQtM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNNaofXvPM7H917I+nceD3VVTnlZG9+lS6fwXg0p29xLEmma1sTWv/M46s+vkWRCYNFrrAfOuPKBUFfrlJS18V127HmAyGgV3ZDyugWsaa3M7mWf4zbBpcBt2ttaaOzJW4QYCpGGldcBWNNamS6jLuRQbAo7uo/mUGwKXUZdGOoqnbI6J12AsoSGLq2puyA0rZXpNOg2lg8ahGUJhqG4YtCwUFfplNVzwAQASg9fd7gh5XULWNNamV3bClDKBRgoXOzaVhDqKp3Seg6YwO593u0NKasDsKa1Ml3T4nC5DcQAl8uga1pcqKukNZDugtC0ViYxNYbx9w1i17YCuqbFkZgaE+oqaQ2kA7CmtUKJqTE68J4EdBeEpmlaiOgArGmaFiI6AGuapoWIDsCapmkhogOwpmlaiOgArGmaFiI6AGuapoWIDsCapmkhogOwpmlaiOgArGmaFiI6AGuapoWIDsCapmkhogOwpmlaiOgArGmaFiLNGoBFJEdENorIOhHJcPa1F5HPRWS78xjn7BcReVpEskRkg4gMrnSeyc7x20VkcnPWWdM0raW0RAv4F0qpgUqpdGf7QWCxUqoXsNjZBrgY6OX83Aq8AHbABv4MDAeGAX+uCNqapmmhllFUihGfkNiQsqHoghgPzHGezwEur7T/dWVbAcSKSGdgDPC5UuqQUqoA+By4qKUrrWmadrSMolKuWpeF0b5D14aUb+4ArIDPRCRTRG519nVSSu1xnu8FOjnPuwJ5lcrudPbVtL8KEblVRDJEJOPAgQNN+R40TdOqtbywBL/VwDXpaf4liX6ulNolIgnA5yKypfKLSiklIg2vfdVzvQS8BJCent4k59Q0TavNyNgoPIaAokExp1lbwEqpXc7jfuA97D7cfU7XAs7jfufwXUBSpeLdnH017dc0TQup9JhI5g/siVVwcHdDyjdbABaRSBFpV/EcGA1sAhYAFSMZJgP/dZ4vAH7ljIYYARQ5XRWfAqNFJM65+Tba2adpmhZy6TGRWPn79zakbHN2QXQC3hORiuu8pZT6RERWA/NE5GYgF7jaOf4jYCyQBRwBbgJQSh0Skf8FVjvH/UUpdagZ661pmtYiRKmTr7s0PT1dZWRkhLoamqadIkQks9JQ2zrTM+E0TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TQsRHYA1TdNCRAdgTdO0ENEBWNM0LUR0ANY0TWuEzNwCXFEdEhtSVgdgTdO0BsrMLWDSKytwRcZ1bUh5HYA1TdMaaEX2QXwBC6Rh5XUA1jRNa6DUqACirAaX1wFY0zStgaLKD3BR2FYi8DWovA7AmqZpDZScnEznsHLail81pLwOwJqmaQ2UlJTE5MmTKSkp2d2Q8joAa5qmNUJSUhLFxcV7G1JWB2BN07QQ0QFY0zStEby5h0mMitcTMTRN01qSN/cw+a9spGNkez0RQ9M0rSV5s4tQAQukYTMxdADWNE1roPDUGMRtgFJ6GJqmaVpLCu8RTfwtAzhQekgPQ9M0TWtp4T2i2VuSf2IOQxMRl4isFZEPne0UEVkpIlki8raIhDn7w53tLOf15ErneMjZv1VExjR3nTVN01pCS7SAfwN8X2n778CTSqmeQAFws7P/ZqDA2f+kcxwi0he4FugHXAQ8LyKuFqi3pmlas2rWACwi3YBLgFecbQHOA+Y7h8wBLneej3e2cV4/3zl+PDBXKeVVSu0AsoBhzVlvTdO0ltDcLeD/BzwAVORr6wAUKqUCzvZOoGL8XFcgD8B5vcg5Pri/mjKapmmtVrMFYBG5FNivlMpsrmscdb1bRSRDRDIOHDjQEpfUWpmsje+y7KM7ydr4bqiromkAuJvx3GcB40RkLNAGiAaeAmJFxO20crsBu5zjdwFJwE4RcQMxwMFK+ytULhOklHoJeAkgPT29QWPytJNX1sZ32bHnASRMsWPPZwD0HDAhxLVquMzcAlZkH2REageG9IgLdXW0Bmq2FrBS6iGlVDelVDL2TbQvlFKTgCXAVc5hk4H/Os8XONs4r3+h7MHNC4BrnVESKUAvYFVz1Vs7Oe3JW4QYCjFADMWevEWhrlKDVaxD9s/PtjLplRVk5haEukq1Wv/O43x874Wsf+fxUFflhBOKccC/B34rIlnYfbyznP2zgA7O/t8CDwIopb4D5gGbgU+Au5RSZovXWmvVOiddgLIEZYGyhM5JF4S6Sg1WsQ6ZpcAfsFiRfTDUVarR+ncep2jmUtTmvhTNXKqD8FGaswsiSCm1FFjqPM+mmlEMSqlyYGIN5R8FHm2+Gmonu4ruhj15i+icdEGr7n4YkdqBMLeBP2DhcRuMSO0Q6irVKPvLzeztfw+W4cKwTBK/nMsZV4a6VieOFgnAmnYi6DlgQqsOvBWG9IjjzVtGtIo+YE/387C2u0BcWGJvn2wycwtwRXVoUDpKHYA1rRUa0iPuhA68FdInjCdv5iosy8Rw29snk4r+eFdkXIOGxuoArGlas0lMjeGKacPYta2ArmlxJKbGhLpKTaqiP56GZaPUAVjTtOaVmBpz0gXeCiNSOzC2bAEf+PY0qHydR0GISESDrqBpmnaScme8zNUr8wjzRzao/HEDsIiMFJHNwBZn+wwReb5BV9M0TTuJZH+5me/634M3vGEt/Lq0gJ8ExmDPSkMptR44u0FX0zRNO4l4up+HZbhoaCdwnfqAlVJ5UnXNIz0RQjul7d+/n2nTprFlyxYsyzp+Aa3VMwyDPn36MHPmTBISEoCfRnlAw7If1CUA54nISECJiIdj8/tq2iln2rRp/OIXv2DWrFl4PJ5QV0drAX6/nzfeeINp06bx+uuvAz+N8njo9YJj8tPURV26IG4H7sJOAbkLGOhsa9opa8uWLVx//fU6+J5CPB4PN9xwA1u2bKmyPzE1hqLShi1JdNwWsFIqH5jUkJNr2snKsiwdfE9BHo+nSbuc6jIKYo6IxFbajhORV5usBpp2ijpSdpADhTs4UnbiJtPRmldduiBOV0oVVmwopQqAQc1XJU07+R0pO8j+g/sIFJSz/+C+BgXhqKio4POPPvqItLQ0cnNzeeSRR4iIiGD//v3VHisi3H///cHtmTNn8sgjjzTsjWiNUpcAbIhIcNK5iLRHz6DTtEY5UlhAG68Hsdy08Xo4UtjwnL6LFy/mnnvu4eOPP6ZHjx4AxMfH889//rPa48PDw3n33XfJz89v8DW1plGXAPxP4FsR+V8RmQEsB3RST02rp8zcAp5bkkVmbgGGr2ob5ujtuvrqq6+YOnUqH374Iaeddlpw/5QpU3j77bc5dOjQMWXcbje33norTz75ZIOuqTWd4wZgpdTrwARgH7AXmKCUeqO5K6ZpJ5OjV7HYdFD4afC+4AqPrvc5vV4vl19+Oe+//z59+vSp8lpUVBRTpkzhqaeeqrbsXXfdxZtvvklRUVG9r6s1nRoDsIhEO4/tsQPvW87PXmefpml1dPQqFusPejFccYgRheGKIyKu/gHY4/EwcuRIZs2aVe3r99xzD3PmzKG4uPiY16Kjo/nVr37F008/Xe/rak2nthbwW85jJpBR6adiWztF7d72PSvfm8fubXo+Tl1VrGLhEvC4DUamdSS8UzT+DrGEd4rGE17/LgjDMJg3bx6rVq3iscceO+b12NhYrrvuOp577rlqy997773MmjWL0tLSel9baxo1/taVUpeKPf/4HKXUjy1YJ+0Etnvb9/znf6djBgK43G4m/vFRuqT9LNTVOuEdvYpFn67R/FDmRQGHAn5OCxhEul31Pm9ERAQLFy5k1KhRdOrUiZtvvrnK67/97W8ZOnQogUDgmLLt27fn6quvZtasWUyZMqWhb01rhFr7gJ1ViRe2UF20ViDvu42YgQDKsjADAfK+2xjqKrUaQ3rEcdcvejKkRxwlpoVy0gcoBSVmwwf3t2/fnk8++YQZM2awYMGCKq/Fx8dzxRVX4PV6qy17//3369EQIVSX7z1rRGSoUmp1s9dGO+El9RuAy+0OtoCT+g0IdZVapSiXgZgKAha4DaJc9V+gvKSkJPg8KSmJHTt2ADBu3Lgqxz3xxBM88cQT1Zbr1KkTR44cqfe1taZRlwA8HLheRHKAUuxbt0opdXpzVkw7MXVJ+xkT//goed9tJKnfAN390FCmQo4EUEohPgsilB5dfwqqy698TLPXQmtVuqT9TAfeRir1BqjcB1HqDRDZgBtxWutW429cRBKAh4GewEbgr0qpwy1VMU07mUWGuxERUAoR0cH3FFVbx9Pr2F0OzwBRgB4wqLVqGUWlPJ27j4yi0A+7igx3kxIfSafoNqTER+oAfIqq7bfeWSk13Xn+qYisaYkKaVpzyCgq5ap1WfgthccQ5g/sSXpMwxZSbCqR4W4deE9xtd56dVJPtndmvrmO2ta0VmN5YQl+S2ECfkuxvLDkuGU0rbnV9uc3BnvWW+XF4CpawQpIba5KaVpTGxkbhccQcFrAI2Ojjl9I05pZjS1gpVSyUipVKZVSzY8Ovlqrkh4TyfyBPfl9aucTovuhKVTk+N29ezdXXXVViGtz4vjTn/7EokWLQl2NOtEdUNopIz0mMrSBN28V5CyD5FGQNKzJTtulSxfmz5/fZOerTiAQwO2uPVzU5ZjmLF/hL3/5S6PP0VLqP/1G07T6y1sFc8bBF4/aj3mrmuzUOTk59O/fH4DZs2czYcIELrroInr16sUDDzwQPO6zzz7jzDPPZPDgwUycODE4I+4vf/kLQ4cOpX///tx6660oZ3zyueeey7333kt6enqNaS1vvPFGbr/9doYPH84DDzzADz/8wEUXXcSQIUMYNWpUcAHLH374gREjRjBgwAD+8Ic/BFvvS5cuZdSoUYwbN46+fftimia/+93vGDp0KKeffjr/93//B8CePXs4++yzGThwIP3792fZsmWYpsmNN95I//79GTBgQDC/8Y033hj8g7R48WIGDRrEgAEDmDJlSnBKdnJyMn/+858ZPHgwAwYMOGahzZaiA7CmtYScZWD6QJn2Y84yLK+JediH5TWb9FLr1q3j7bffZuPGjbz99tvk5eWRn5/PjBkzWLRoEWvWrCE9PT04PfnXv/41q1evZtOmTZSVlfHhhx8Gz+Xz+cjIyKiyhNHRdu7cyfLly3niiSe49dZbeeaZZ8jMzGTmzJnceeedlHoD3H7X3dx256/ZuHEj3bp1q1J+zZo1PPXUU2zbto1Zs2YRExPD6tWrWb16NS+//DI7duzgrbfeYsyYMaxbt47169czcOBA1q1bx65du9i0aRMbN27kpptuqnLe8vJybrzxxuBnEQgEeOGFF4Kvx8fHs2bNGu644w5mzpzZFB99vdU2EaPWkQ5KqWNT7WtaK5dRVMrywhJGxkY1bXdF8ihwhdnB1xWG1WUkZfnF+DHxFLtoG98OI7z+2dCqc/755xMTEwNA3759yc3NpbCwkM2bN3PWWWcBdmA988wzAViyZAmPP/44R44c4dChQ/Tr14/LLrsMgGuuuea415s4cSIul4uSkhKWL1/OxIkTg6+VlXvZkV/K6pUr+Mf/vUGpN8B1113HtGnTgscMGzaMlJQUwG6lb9iwIdiCLSoqYvv27QwdOpQpU6bg9/u5/PLLGThwIKmpqWRnZ3P33XdzySWXMHr06Cr12rp1KykpKaSlpQEwefJknnvuOe69914AJkyYAMCQIUN499136/kpN43aOlwysUc7CNAdKHCexwI/AinNXjvtuNZteouM7E9JTx3DwP7Xhbo6J5T17zzO7mWf02XUhZxx5QPHPb5ZxwonDYPJC4J9wN6ofpR4D2K4fHjNMIwjHtqGN821wsPDg89dLheBgJ1z4sILL+Tf//53lWPLy8u58847ycjIICkpiUceeYTy8vLg65GRx69TxTGWZREbG8u6deuCr+0/XM6+w/b5lDPluk0N5SuOeeaZZxgz5tgMCF999RULFy7kxhtv5Le//S2/+tWvWL9+PZ9++ikvvvgi8+bN49VX675ge8XnVPEZhUJtoyAqRjssAi5TSsUrpToAlwKftVQFtZqt2/QWf3//UX5cuJq/v/8o6za9dfxCp4j17zxO0cylqM19KZq5lPXvHH8Zw2YfK5w0DEbdD0nD8MsR2kQUER5eRpuIIvxSv4xkpd4AynmsixEjRvDNN9+QlZVlly8tZdu2bcFgGx8fT0lJSaNu5kVHR5OSksJ//vMfwA6m2Vu/Q0Q4fXA6iz/+gMhwN3Pnzq3xHGPGjOGFF17A7/cDsG3bNkpLS8nNzaVTp05MnTqVW265hTVr1pCfn49lWVx55ZXMmDGDNWuqzhXr3bs3OTk5wff8xhtvcM455zT4/TWHutxyHKGUmlqxoZT6WET0opwngK1L5nPHwh6UtOvFBRu2szVmvm4FO7K/3Mze/vdgGS4MyyTxy7m0H5ZHTk4OycnJJCUlHVNmZGwUHic/g0eEOLeLp3P3NX13BOAOC1DR6BJnu65KvQF25JeiFOzIL8WoQxDu2LEjs2fP5pe//GXwRtSMGTNIS0tj6tSp9O/fn8TERIYOHdqQtxP05ptvcscddzBjxgz8fj/XXnst9//+YR6f+QR33zaF1194kosuuijYRXK0W265hZycHAYPHoxSio4dO/L++++zdOlS/vGPf+DxeIiKiuL1119n165d3HTTTViWnUv5r3/9a5VztWnThtdee42JEycSCAQYOnQot99+e6PeX1OTijueNR4g8imwDPiXs2sScLZS6oTNkpaenq4yMk7+VZO++cMf2LDv58Egc3qnrzlrxoxQV+uE8OHM18nd3hnEBZZJh9N2kOXfj2mauFwuJk+efEwQ9uYeZt7bK1geZ9HFF8asfh3wq+q7I9LT02nMv7FAoJQjR3ZQ0csXEZGC2123IF/xtb6if7BTdBsSoo/+Yn9iOXLkCG3btkVEmDt3Lv/+97/573//G+pqNUh1v3sRyVRKpdf3XHUZBfFLoCPwHvCu8/yX9b2Q1vRcqWOxDBeIC0tcuFLHhrpKJ4z0CeMxXAAmhhui+6RgmiZKKUzTJCcn55gyq5ZtJKd0BYk7V5DZZi8+y6rSHeHNPczhJXl4cxufFNDtjiQiIoXw8E71Cr7wUyY1gVaTSS0zM5OBAwdy+umn8/zzz/PPf/4z1FU6IRz3N+eMdviNiEQqpUKfRkoLSj63H+syM7FMheFxkXxuv5DUY/e270+4BO2JqTFcMW0Yu7YV0DUtDr/nMGs2rgy2gJOTk48ps75kL3ustuxV0bQ5UIyRrFAiuERILxfy/7URFbAQt4EKNHwJoQpud2S9Am+FikxqFTmEWyIAP/roo8G+3QoTJ05k+vTpNZSoatSoUaxfv745qtaqHfc3JyIjgVewU1J2F5EzgNuUUncep1wb4Csg3LnOfKXUn0UkBZgLdMAeaXGDUsonIuHYKTCHAAeBa5RSOc65HgJuBkzgHqXUpw15syebxNQYLr9/SDDIJKZW36/WnE7kRToTU2MqfSYx/HLMVezYnEVK357V9gG7k1P4NCcME8FVqBj01RpK2sTQ2XeAuNMvwApYiAIrYDVJAG6Mls6kNn369DoHW63u6vIbfBJ7VYwFAEqp9SJydh3KeYHzlFIlIuIBvhaRj4HfAk8qpeaKyIvYgfUF57FAKdVTRK4F/g5cIyJ9gWuBfkAXYJGIpCmlmnb0eitVNci0vOoW6WyqAFzRsna3O8AR33d0TrqAngMm1HpsTa1wb+5h2iw4SJ9ADJJ1EG/CYcJ7RFc5xgyPxmIXIFgovvN2Qnkhj278d+0irlL9cQMBBQGRY66hVa/UG2jR1nprUqdPQymVJ1X/wR03+DkrKleM4/E4Pwo4D6i4VT8HeAQ7AI93ngPMB54V+6LjgblKKS+wQ0SygGHAt3Wpu9a8mmuRzoqWdVhcET0vyUXCFDv22KMfjw7CNbXCM3MLgsvA98ousVutClTAwptddEwADt8yC7cxDNNyIQKmAnBhYZKNj/s4whm42ICJ9zg3rzXbTyM27JU/dPL5qurySeQ53RDKacn+Bvi+LicXERd2N0NP4DngB6BQKVUxbmYn0NV53hXIA1BKBUSkCLuboiuwotJpK5epfK1bgVsBunfvXpfqtXq7t+kKkJoAACAASURBVH3Pd19+AUC/c84LyVf/5lqks6JlHdW5FHEpxABQ7MlbdEwArq4Vvic8kete+ga/qfC4hDnjBlJkWKwx/Qw2PJxbzbeG2LKvmTZ4FVsLexHpKWHu1isxLXAZJr3jtvPVod5sDvjwuA3C3Y2fxd+YlqGvvAxfWRlhbdsS1qZto+vSXEq99iQQBXrtu2rU5ZO4HXgKO+jtwp6EUWv/bwWnm2CgiMRij6Lo08B61uVaLwEvgT0Mrbmuc6LYve175v3Pw+R3jGBvTCJdXlzJ7bc/FLIg3NTXrWhZl+6JRJn5gEJZQuekC2o8tnIr/JXlX+EzBYULv2nyr42ZLJIwfFiEiZ83MRly1Hlc6kxS273LabE5KEvoGrGXbYU9SYv9gWEdbuLc00YEW9S3fdy4ANyYlqGvvIyiA3mk9BnIji3riOmYdMIG4chwN7gNcNlfKXTwraoun0ZvpdSkyjtE5Czgm7peRClVKCJLgDOBWBFxO63gbthBHecxCdgpIm7shPAHK+2vULnMKSvvu43kd4zg7YirCHhduNuYpK5ezI0nyA2wxqrcsi4tz+Vw+Sa6dDm32j7g6lrh3da8gdsYGmzBHindjc9MxgL8psWK7IMM6RFX5TxnXjGdhS/F0KbDBsoPnk7fYZ1JCl9G5w43Mfg8u9fs6DIN1ZiWoa+8iPBoLwiER3vxlRedsAEYl6Ai3fYC0GJvaz+py2/8GWBwHfZVISIdAb8TfNsCF2LfWFsCXIU9EmIyUDEae4Gz/a3z+hdKKSUiC4C3ROQJ7JtwvYCmy+XXSiX1G8D+9asIlLtQuDAt+EF5Ql2tJtUl7WeYbaOYM3sbppnKpm276JycV+0IhqNb4Wf3S8fnes5pwWbRpcNdLM838AcsPG6DEakdjjlHYmoMl9z666NGlUxpsvezbv86MvZlkN4pnV4x/Ru8KrLhsbAq7sIIHCkvZvz5kygoKMDv9zNjxgzGjx9PaWkpV199NTt37sQ0Tf74xz9yzTXX8OCDD7JgwQLcbjejR49m5syZ5OTkMGXKFPLz8+nYsSOvvfZak3TllZgWFd3lStnbke6mSTp0MqgtG9qZwEigo4j8ttJL0UBdPsHOwBynH9gA5imlPhSRzcBcEZkBrAVmOcfPAt5wbrIdwh75gFLqOxGZB2wGAsBdegSEHXAuPfssliwyg628CwedEepqNbmsdVvt/l0BMxAga93WagPw0Spayv3yFtE56U56DriCrqf9dFOuppZsc40qWbd/HVM/m4rP9BHmCuPl0S/TK75/g/qAw8JjCBypSEYoRMck8N577xEdHU1+fj4jRoxg3LhxfPLJJ3Tp0oWFCxcCdmaxgwcP8t5777FlyxZEhMLCQgDuvvtuJk+ezOTJk3n11Ve55557eP/99xv9vsOVhX3vvfK2VqG233oY9thfN9Cu0v7D2C3UWimlNgCDqtmfjT2K4ej95cDEo/c7rz0KPHq8a55qxp13BbGdVrBs6w+M6n0aZ/cbEeoqNSlv7mHi89wYGFjKwsCgsxlb47He7CLCU2OCoxt6DphQpctiSI+4JutCqK+MfRn4TB8WFn7LT8a+DAYmDGxQn2jFLLqKKcxKhfHww/fx1VdfYRgGu3btYt++fQwYMID777+f3//+91x66aWMGjWKQCBAmzZtuPnmm7n00ku59NJLAfj222+DKRlvuOGGKoncG8Pl9xF75Ag+l5swM4BLIqBN+PELniJq/O0rpb4EvhSR2Uqp3Bask1YPZ/cbcdIFXrADav4rG4kLuEiJGcHyDhZnFbpIGZRW47EqYLGxvZvvL+zC2SnxTZ5A56UXn+LbHWWcmVL//tb0TumEucLwW348hof0TvVOG1CFPYNOcLsjmT17NgcOHCAzMxOPx0NycjLl5eWkpaWxZs0aPvroI/7whz9w/vnn86c//YlVq1axePFi5s+fz7PPPssXX3zRqLrUJjw8HE9xMR6nz6Ryqkytbn3Ar4jIRKVUIdhL1WOPyz1hk/ForVPl3MapB0ahAhYbog2mp3fEbwgfGULvWBdHhy5vdlHw2DsGtcFfUMDTRYXHJNCpbrJGXadRv/TiU/w9JwVTDL7KsfAUF9frvQ1MGMjLo18O9gEPTBhYr/K1KSoqIiEhAY/Hw5IlS8jNtdtLu3fvpn379lx//fXExsbyyiuvUFJSwpEjRxg7dixnnXUWqan2+rojR45k7ty53HDDDbz55puMGjWqSeoWFhZGfHw8Xq+X8PBwwsLCmuS8J4u6BOD4iuALoJQqEJGEZqyTVkdFRWsoKFhJXNxwYmJqvSd6wqvIbdwrT7E4KYPpZ/0Pce7OZHZw4xewBPzKTopzdMs2PDUGcRs/HctPCXQqjt297Xve/st0rIAfw+3hmj/ZPVp1nUb9TbbCNAzsWXIG5YH6j3QcmDCwSQNvhUmTJnHZZZcxYMAA0tPT6dPHHu25ceNGfve732EYBh6PhxdeeIHi4mLGjx9PeXk5SqngskTPPPMMN910E//4xz+CN+GaSlhYmA68NahLALZEpLtS6kcAEelB5V51LSSKitawZu0NWJYPwwhj8KA3WnUQPjq38Xcxb3He9ffRd9dawmQ4fsBjCCNjo44pG94jmrDrrVqPXfHNKgJ+PwaKgN/Pim9W0TW2bZ2nUfeNieXrYoUFGChcRuiXU6xYVDM+Pp5vvz12YmhycnK1K0usWnXsIKIePXo0a1eEVr26BODp2HkcvsQeyTcKZ8aZFjoFBSuxLB9gYVk+cnI+44wzWm8A7ravLxv6/5TbuH/Z+2zaczux+HiIvhR1+TPnJfautl+3qGhNlWMPqGmcl9ibgm25/GPzNgb2TWNXmy6Y4gJlYoqLXW26MKJftzpPo578yys49PRicj1+evg9fBYZ0Zwfh3aKqEs6yk9EZDBQcafnXqVUfvNWSzueuLjhiHiwTB+WEpYs2U379tWPkW0NXKlj+dHyktMpnOS9XvqdFhv8A9NTbea08JUk19DCr/zHqKe5mRE/fIj/iyI+ca0BLL7M3kDKmWNZ1+0ahvp9rPaEcefIdLr0iKvzNOrE1Bjuu+f84BjhxV8/1iyfg3ZqqW0ccB+l1BYn+ALsdh67O10Sa2oqqzW/mJjBuIzfsSN7IYWFnSgp6UBOTk6rDcCFvbvyZvdD+AVcfdvSt+0oOpW9i4iglODz1bwGrM+XgmlK8Njygm7sVwUIFiKAsijLyeWRNkkot+ISt5DgDGWvzzTqUGee004+tbWA7wemAtWlrq/IaqaFUGrqBSz7aletScZbA2/uYVZ/v49AigdlCMqAtWY83TaOJjpmD4eLOtO2TSRpx45AA2Dfvkg2VRxbmEhaSSydJQ6FgVIWFgZnRCUipoUAmKrabGha87C8JsprIuEujHA9C66y2sYBT3Uef9Fy1dHqI8GKYaxvELvVIbqY7UmwWmfrzJtdxJB8P55kDwFnSfjzO3dk7ZFEios7HvePS3JyMl9+6RxruOiZ/jNSBqVRsD+VdU4f8LCEbuRv/WlFi/AGtGQzikpZXlhS7Y1ArXqW1ySQX2bPQxbBHd9WB+FKauuCqD7ztUMp9W7TV0erD292EQmBaBJUNAjN0qprzuWGNiz6mG0rl9O398/BiOKSXT7EECYNS2ZkWgKnT55c6yrGFZKSkpjsHBtlQNH+bA56I7lwaF8uHNo3eFz8LQOOmS1XVxlFpVy5Zit+BR6BGEsPBKoL5TWpnAxCeU3QATioti6Iy5zHBOycEBVjVH4BLMdeoFMLoYrxr41p1dWmOZcb2rDoY9a8/h4Jbbrzn0Nf88pF4wgYLtyWxaW+A0ACSUlJJCUlkZlbwIIlWbXmcEhKSsJVVlJrfcN7RDf4D9QHW1biVzFY4sKvAhzxext0nuY0duxY3nrrLWJjq07XfuSRR4iKimLatGktXicJd2G5vViucgyzDe5wPXqkstq6IG4CEJHPgL5KqT3OdmdgdovUTqtVeI/oRrXqjqdZlxta/h0d+lxPRocw9oZDwHDBYT/qYDn/OXiQiwf0I6OolP/sOMA7n2dhHfIS5jZ485YR1QbhjKJS3s7aiT8xhYTSItxlpXWqb15eXt1a2QVf4Q67mIBSuDFxWydWAFZK8eGHH2KcAOOTK7Nc5fgi9mLfNhLcrrYYNO0U8dasLuOAkyqCr2MfcGosOdEKNKZVd7T17zzO7mWf02XUhZxx5QM1Lje0bNZz5K3dR9KgTpw26rwGdVEcSUnnvs7t7FEPClwF5bjWHAILlhmKNwfsZnr+AbyWQg1qT9jqg8hhX7V5fDOKSrlqXRa+tvHImF9y2fpvSCw6SFhC5yrHHT1zMC8vjzlz5gRvYk6ePLnGIHxu19Px7f0fthh96WNt5iVX/VN/Hlm7liOrVhMxbCgRg47JU1VvOTk5jBkzhuHDh5OZmcnmzZs5cOAA8fHxPProo8yZM4eEBPubxJAhdvr51atXc/PNN2MYBhdeeCEff/wxmzZtwjRNHnzwQZYuXYrX6+Wuu+7itttua3QdTbOUn+ZtKUyztEErQZ+s6hKAF4vIp8C/ne1rgEXNVyWtruraequL9e88TtHMpah2fSlauZT1wBlXPnDMONlls55j04rTsIw0Dn29k4xFv0dBvbsocgak4D94CMsQxFL0PVTMdstuJZmWxX/XbsXfNc7+ryuC6hCG50ig2jy+ywtL8FsKC0HEYHdsPInFBZRUynxY3czBnJxSTNNEKYVpmrUO46vIqjYgbxGdk27h1bD6jQM+snYtP940BeXzIWFhdH/t1SYJwtu3b2fOnDmMGDEieKMyMzOTuXPnsm7dOgKBAIMHDw4G4JtuuomXX36ZM888kwcffDB4nlmzZhETE8Pq1avxer2cddZZjB49mpSUmof/1YXLFQlO3mNE7G0tqC4TMX4tIlcAFSshv6SUeq95q6UdT31ab3WR/eVm9va/JzgTLfHLuZxx5bHjZPPW7sMy0kBcmOYeLMuOcvXtojg7JZ6nCwvxK4XHZXBR/E5ycuIwLRcuw2R4xA5WGe3BUrhcBpN6JXLVmIRqux9GxkbhMQRlWoiy6FKYDwo8ZT+tvlB15qCfgoKVJCdfisvlqvMwvqrpLesZgFetRvl8YFkov99uCTciAPt8PkpKSujevTsjRlTNhrds2TKuuOIKIiLs/tZx48YBUFhYSHFxMWeeeSYA1113HR9++CEAn332GRs2bGD+/PmAneBn+/btjQ7AbnckEW1TMM1SXK5I3fo9Sl2Tka4BipVSi0QkQkTaKaXqlw5Ka1I5OTl1br3Vhaf7eVjbXSAuLLG3q5M0qBNFK0wsAZerM5ZhBFvA9VkROT0mkvmDegaHdfXYe5h21qNsKUihT9wOrug3nXMSf3q9ttSS6TGRzB/Yk9e/yMCXvY7E4gIASvYEc0gRFzccwwjDsvwYhsfphvhp9ERTfIuoTcSwoUhYGMrvRzweIoYNbfC5fD4f+fn5lJaW0qZNG3w+X6OT3SileOaZZ6rNHdFYbrcOvDU5bgAWkanYuR/aA6dhL875InB+81ZNq01ycjL5tGN3IJIuntJGT8JInzCevJmrsCwTw21vw7HD0EbdfBfg9AEP78Rpo/7e4GFq6TGRPwXWmFFcwXQO7vqGDl2vo2PvUXR0jqnrueK6dWJB9naiu9mTN7p36srhJXmEp8YQ02Mwgwe9cUz2uIqRFs0tYtAgur/2apP0AXu93mO2Kwfgs88+mxtvvJGHHnqIQCDABx98wG233UZsbCzt2rVj5cqVDB8+nLlz5wbLjBkzhhdeeIHzzjsPj8fDtm3b6Nq1K5GROnA2p7q0gO/CXsFiJYBSartORxl6+60oPvP1wWdabFIGN1pRNCaMJKbGcMW0YVXWQ6tpGJodhH/SFCMj8vLyyNkPyb0m0bGBATHnSAb9B3yKGBZYG5G1Qzhc2BNxG8TfMoDtsb1ZLl0ZSdQxOYVbQsSgQU3S7xseHk5xpXzERyc5Hzx4MNdccw1nnHEGCQkJDB36U2t71qxZTJ06FcMwOOecc4iJsYcu3nLLLeTk5DB48GCUUnTs2LFJliTSaleXAOxVSvlE7P40Z8ViPQo9xL5ZsxufqVAIflPxzZrdjV5u5+hcB/UdhlZ5dMTRQbo2Ff3ZgUAAwzAYO3Ys6en1C5GZuQWsWr+AoX0txACFxZHY72lb0BMVsFiRnc+v5LDd5yzCO4N6NfmKGS3FrVzEqAgiuvUkc/EK3Mqe2JCTkxM8Zvr06UyfPv2Ysv369WPDhg0A/O1vfwt+zoZh8Nhjj/HYYzrJUEuqSwD+UkQeBtqKyIXAncAHzVst7XgiUVgGYIFp2NuNdXR3Q1K/AfYNbBQiBPt4583/iCUbcvnF6T24+qqxAFVGRxStMCkunEGH1LRauyYqRnEUFRURCAQAsCyLjz76iL2rvubw5kMkDepExAXXHXcxzRXZB1lb2I8h1hoMFMoSIgp/BgLiNvgi7Aher30n3mtZfJSzi/QzakgucYJTXhOPcuFxEgrVZ3bZwoUL+etf/0ogEKBHjx7Mnj27GWuqHU9dAvDvgVuAjcBtwEfAK81ZKe34ilLaEbDiodCHxIZRlNLu+IVqUV13w4GNH9B7V4D8mHjii/I5sPEDvt6wg4dX+zClG5+v9gEfcfVVY6uMjgiYe/h+1UokY1WNw9Mqj+IwDMPJZGb/EbEsiy3b29HW34/NGQHmZX1NAKPWiRgjUjvwzBfDmLUCBsZ+x7AzxtH92nHBSSoq5zugjTMtVggUHgSqX8zzRCfhripDu6QeU3uvueYarrnmmmasnVYftQZgZ0n575RSfYCXW6ZKWl0Eh3G1D8ctwtkp8Y06X3XdDbu/3MzeNHtoWmknE++Xc/myk4EpaSgMLFF8uXUzVzOWpEGd2LX9e3a3KyexqBR3qUJZqsaui8qjOCzLonfv3mzbts0OwpaF2x8L4iLXY+K3BCXgD1jVTsQAe8Xj5y/swMebU9nb5xJiBv6M8JjIYFB1F3aAgyUghn3+2A5VFvOs6CduDUHYCHfhjm+rM4ydBGqdt6iUMoGtIqJnvp1gKoZxXR8Xyy9VG6TQ16jzJfUbwL6ILmTEDmZfRBeS+g2wh6YZFUPTXHi6n8fw3sW4DRMD0x6v29u+GbS9TXteMpJ4v7gPr7gH4u+eghhGjcPTkpOTcblciAgul4uzzjqLm266ifPOO48hneII90aCZdLDr3CLwiXgcRvVTsQA+OKrVXz47hv8O7U380yDK9duJ6OoNPj62OSuhImBKEWYYTA2uWtwMU8UqICFN7uoUZ9hSzLCXbiiw3TwbeXq0gURB3wnIquA4L9opdS4ZquVFlR5peCB/a+r8trmF5+jx9ZDZLU1mNR+VI1fz+H4N8j2hCfyfuI4/KbFGpfB1eGJ1Q5Na9vhB8rkIbYWpNA7bgcX9/8Dh5fksX73TgJWbxQuTAt2duzKDSNH1dgHnJSUxKCrr2PxngOc37ljcChYUlISjBrF+089yervjrCuTSzKSObaod2ZMLhbte8vM7eAmf/5mpLOp2OVmKi4sGMW5ZRCH2EZ+agYD2FFfiS1R7MnM9K046lLAP5js9dCq9a6TW/x/Ntvkrq/F88nvMmknQdoU9aZpH4DWPXxJ5RmnY7yuOjlMxl76CtWZPeqNkAdfYMMnjsmCK/IPhiczhuwFCuyD3LXL3pWGZpmBXaz5YssBiVcS+8OW+jY5naKFvxAafSH9HKF4zZMTAtchkn3dm0ZfsXVNb63jKJS7thdjF+FM393Me92Kq0yKmFV9wt4a8+P9obdSK3xj8s7a3ayJbIn1mHBs/oQZp92uLpHVV2UM/sg1iEvroNelNjbQ37Rs1mTGWna8dSWD7gNcDvQE/sG3CylVKClKqbBpoVfM2jv3ViGi9N37iRjtr1Ej8vtpm1Ydyx3/+DMtZ5lFsNr+Hpe+QaZJfb20UakdiDMbeAPWFW+6md88ho5Ow7RcYOHvRvX4nWHU5aUBkZbYqK/oG3KNrs13H0H1xpF7PS3p6Pl55eX1p7I5T87DuAzLTAEn2nxnx0HkDhfcLSDHHX80dtHv2ZhLxkvSuHZUsSNiTuA3jydu4+RsVGMSO2A27CH7LkMCb6/pkxmpGn1VVsf8BwgHTv4Xkz1SxNpzSguv3ewD9Y099gJrZ2bZJ72CsMywTIxlElU7/bsXvkFT95/Px/MewewRzasfG8esSlhwWPN8vWUmj+yYdHHVa41pEccsy8dwF09E5l96QCG9Ijjw+f/Hxl7D5MfYfB9iYnX5SGsXWeUCAoo8viZueZ23vthLDPX3I67y2nEDbuIc66YctzZZUaBF5dlIpaFyzI5+ONhJr2ygn9+tpVJr6ygX5cYwtwGAoS5DSYM7lbjuSYM7obLXvwNsEcHvLumLRPWZvH37D1ctS6LLSVlKNNODq5Ms3G/mBNccnIy+fn1Xzd39uzZ7N69O7h9yy23sHnz5qasWrVO5bHHtXVB9FVKDQAQkVnAqpapklah33kj+fHdg8G8C8rlxlIWLrebMVNvZ9XHn1CwOZ+4fvG4Op/G1ndeA2Drzq2UHdzPnuWfB4eV9RoIuzbv4rA3D78XPn/5OQBOv+BiwB6OFfbhetLUIcJy9uJNjCZnxyGIMILNz0BEND2sOIpRWMpia2EaAcuNwiBgCa8UnY7ZMZx39xTT6aguhaMN3/AR+Xv6szVK6F2iOL3NKr4IDMRS9miHgiM+/j11xHHH/4L9x+O353Rl5tI8FGAIGHEu/ErZo6MtxVtfZBBQLpRAQAlz5y1kyO+ub4pfU53tzS6qMtPwRDN79mz69+9Ply5dAHjllZYZbfrYY4/x8MMPt8i1TjS1BWB/xROlVKBiJpzWcpIvHsrFrObHVTl0H9afsNNGVJkocXmlm1tP3n8/YMdKBexcsxIqDSvrkJrGkXIvhzfkBctsW7k8GIB3rN3GQiMTC4u17CCwfCed09bhLepOcXE8KIPTkruTltgf/95IlsdZRBabWGXYfbSGQaB9GxQccwOsOlt+8LOkrQ+z2F5uO/mgi7DYql0gQ3rE1Xl2X3hEKf6hHZA95VjKYkj8Yj4lBQsDjyH03LebrSRhKnABHfLsll5mbkGdgnxj7c0u4r9PrsUMWLjcBuPvG9QkQfhf//oXTz/9ND6fj+HDh/P8889Xef3yyy8nLy+P8vJyfvOb33DrrbdimiY333wzGRkZiAhTptjfWDIyMpg0aRJt27bl22+/5eKLL2bmzJmkp6fzySef8PDDD2OaJvHx8SxevLja+jzyyCP8+OOPZGdn8+OPP3Lvvfdyzz331FjX6dOnU1ZWxsCBA+nXrx9vvvlmoz+T1qS2AHyGiBx2ngv2TLjDznOllNIdZy0g+eKhJF/801z+mmaV9Rw+kq07twbnw3UbPLxKCzip3wDatmtH7oa1wTJpw0cGn+9xFWJhoQQi2+2jKP5fxHQ0iSwoY+OG7gxuH801993O8q37+cOu3fgAlIkqCmAU+LBiPUisBwPwGHLchSsPJnXBzAclYCrYnZjIhL7dEKhxtENtdsd0xMo/hGd3OViKL/acyYT0TE4bMo6RsVGUZX1Nl7Uu9oYJiT7FyEGRZOYWMOmVFfgCVq2TPJrCrm0FmAELpcA0LXZtK2h0AP7+++95++23+eabb/B4PNx5553HBLBXX32V9u3bU1ZWxtChQ7nyyivJyclh165dbNq0CbDTVMbGxvLss88GA25lBw4cYOrUqXz11VekpKRw6NChWuu1ZcsWlixZQnFxMb179+aOO+4gKyur2rr+7W9/49lnn2XdunWN+ixaq9qWJNIDDFuRy66+EoCslcvpOXwkl119JbvPP7dKi7kieG9buZy04SODrV+AngN78826FZimSUzsPkRMfjiczD/X3olpuVlxyE37f77BD1E98fVsYydSVwYS48GK8WAoi/Hbt9Dn4guOmz4S4NqrL+G9F74moARDFJ+WdsRc+SMel9Ta31uTLhtX4CrrAZayWwiW0NZM5p4enQDwXnADPbLWIhaocOh0wQ28kr0fX8AKdnvUNMmjKXRNi8PlNjBNC5fLoGta46+zePFiMjMzg8l2ysrKSEiomifr6aef5r337PTdeXl5bN++nd69e5Odnc3dd9/NJZdcwujRo2u9zooVKzj77LODuYHbt29f6/GXXHIJ4eHhhIeHk5CQwL59++pU11NRXfMBa63AZVdfCU4ghmOTqYPd51s58FZISkpi+EUTWLd5Gx3jE1DWOrYe7BXs4/X6TTJ3JXB+RwuPgoClEKX4RcZGytq1o8f+MC4bfQ7nOgHveLbuLcYEEGU/Oond/aY9rKy+gTBiRRYXtevCYlEoBW4UE4f91JLzZhdhYPdni7M9IrUDHhf4TYXbJTVO8mgKiakxjL9vUJP2ASulmDx5Mn/961+r7K/I77B06VIWLVrEt99+S0REBOeeey7l5eXExcWxfv16Pv30U1588UXmzZvHq6++2uj6VKicnc3lchEIBGqs66lOB+BWLqOotE5Jy48nM7eAuxfk4gt4CHMn89wFjxBj5KGcO3BKgTIV/QtNnlt9hMw4AzPrU9rlZaJw8XWn8XznK+fcOtb3i7VrsAfhuICKtYMMBJOCw3lA3ZO7g50ofvCKAB09Hna6TM7tvbdKEK9u0kXP2CymDXmW7w8m87MOOfSM7Yw976h5HJ1trrHOP/98xo8fz3333UdCQgKHDh2qkqayqKiIuLg4IiIi2LJlCytWrAAgPz+fsLAwrrzySnr37s3119s3I9u1a1elfIURI0Zw5513smPHjmAXxPFawf+fvTePj6o8//7f9zkzkz0zkz2BrISEsEOCRJDNfaFa0aptbcW1Vmxr/fbXR7uoba3deNqf/aq0VusKtSqCK3UBURYJZGENhISQELKRZTLZM3POuZ8/zmRIIOxB0c779QqZOWfmPmfOhM/cbSnMnAAAIABJREFUc93X9blO9lxTU1OxWq14vV6s1lPvs/dlJyDAX2L6m1F6DYlVEbw+OfO0RXhTZcugr+N7jHxiMvIR5b64soBWw2Bjh4fYbkH0/o8I7W0lwT6dht4agrXaE84gB56vEhdH0IFDGAYowkBizuhURSc1ogK48pTOv98o3r2ti/jkJCZ97bpB+4fqIF1fVUCGvYIM+15AxeUq8Bu1fxkYO3Ysjz76KJdeeimGYWC1WnnyySf9+y+//HL+9re/kZOTQ3Z2tr91UW1tLbfeequ/nVT/rHThwoXcfffd/kW4fmJjY3n66adZsGABhmEQFxfHhx9+OCznmpqayl133cXEiROZOnXqf90inOh3oPoqkZeXJwsLC7/o0zjr/LW6kT9U1qNjziP/T0aiP+Z5qhRVu/jFq8vIiNxLZXsWj95glj1/+5lNeDUzbnlB51by2t1Uhim0ZczEGRrKCHcLM1otbB4RxL0/ODq0cazzVZDMa3mHOFct2c4KpITytkxGOyrIiluII+0qf3bCeNSTqlbbVHCQhW9uxyPlkItq7//z17gK1uCcfiGX3fbQgEadZpuiqVNeOmkBzsvL47/hbyzA0Qz13gshiqSUp+zzH5gBf4npb0aJbwZ8osyD45Hp2M9P8p5ESi9CfEim43zs9qksveNwLi6YtycnhfOzg/V4hUQ1Irhm22fc5DyxcA08X1UazLIXkOncjTAsFO2ah8sZT80uL+E9wSz6yJedoCo8LkMZZyjHdSzrq27n4zfL8BgSg6MX1d7/568J+ccm9IhJhGzfxPv8mstue2jINkUBAnxeBAT4S0x/M8ojY8Cn43HrchWA9CIwQGoUlrzOzqICUmJzGO11YA0NYeL0EeSmOvlrdSOaACkUdAEHHVFo9p5TOt+4nUuJDdoDAspFJk9MuB0vKpYknSurWvBouj8cUoyXcQT5HcuGFOBKN1MMBStmArtVDF5UO1Swl84BXZ/DC16G28BunxoQ3tPgueee4/HHHx+0bebMmYNCIAFOTECAv+QMamwJp+1xO7BrMKjsfmMLEW0tTEnIRoh2vOXtbAdSxzSS3rMLqzIej26gSoORHS4yL519Sue7qz6dtR3Z7JFjaSEWTaimoCNoam3AKuLQpMSqCtLoY6uoJ0lEEXuMRaygDDsTrDYe90KJYjDvmuxB4QdL8oUYNYe7PluSh+76/GWhS9Pp1A3CVYUwy+efMXrrrbdy6623fu7H/aoREOCvAANbCYXXhh/lcXsyAmy3T0WMfon1TTV4SyqxN65hfPQ4FKGiCIFEsmHbRl7Y+QlZzjIedITwmfU+DCOcK6+8kuTk5FOqKvvME8fvlF+hCwUFA2EYCAGKlMzrU7jTCKYEnTT6qAgqQTd0VKWKWGUiyRz9evoX2WZUupk3xMx/3reu443Fm5G6jrCY97+sdGk6+3r6+htiMCok6AsR4QBnzlkTYCFEMvAiEI9ZHfu0lPJxIUQU8G8gDagCbpBSuoRZ6/w45vJ3N7BQSlnsG+sW4Be+oR+VUr5wts77XOJkmlwe2Uro+u89csoet4XuLl5taOWVeoEmU7CkJXNrspt0ZQK99n30RO2homkCf6ly4DUuxaJcxA1T32KNMxxNWPm0rp3fygYefbnkhFVlbncxLlcBO/pGoashSKFgSMhpqCK8r4cRbc3E9AUjnZ8xx3GItrY4ai2R1DliSGxrZsPW3dx0DKOf4zmbJWTYWXBE1+cvK526WVEHZnpgp24EBPhLytmcAWvA/0gpi4UQEUCREOJDYCGwWkr5eyHEA8ADmH3nrgBG+36mA0uA6T7BfhjTmU36xnlLSuk6i+f+hXMyHr5wdCuhg017mHzH5ScdAy50d3FdSTmefuMaBBoG3bOm01u6n9q8PyGFl41qF9q+K5Ao6AYUt01Bi1KRQtBnSF7ef5AQqxt7jMTrFqwr2UP3AY20tDS/M5rbXUxx8c0Y0kuOyMFqPIwmJBYJYw7VEt/egkDQENVE9MQP6FF0Dhg5vKM8Ys6UDYPIHp2bTvOa1qoGm4I08lWDhAHX+XQ6OX+RhKvKwJZwhKvHbWwT4BzmrAmwlLIeqPfd7hBC7AZGANeAP1//BWAtpgBfA7wozby4TUIIhxAi0ffYD6WUrQA+Eb8c+NfZOvdzgZPx8AWzlZBqsQzyfDgVj9v3qmrxGAZSKCAlAgMLGjNH1NHbuBcpvKBIspx7sSqX4zXMXN3kqDq2tmmorV0YUUFstVtRzs9EKgLVMNi7cxPuba2oqsott9xCcnIyTfs/wdC9oBiMZje/q11NRe9sJrXq2KfMxBLWRcXmEnT7dhRFRwgoVXLQhTDjwwooyafX+24o34fuj5ad1IfcuUaYRWVUSNAXGgMOMDx8Lh+dQog0YApQAMT7xBmgATNEAaY41wx42kHftmNtP/IYdwkhCoUQhU1NTcN6/idDv/du3d7dwzJe8pR4FEOnVvFSEOSld8zQbfmSsnL4xi9/y8wbbmb23T9mX2MzNTU1Qz52yOe7m1ANAyENLHi5kA/5uXiMeaOnoKROQUoVDIXR9oM8dUU4czLhx1P/RpcMw1bYjKWiA1thC8LtxVBVpKKiKwoHI6OQUqLrOlVVVQCEtuYgpAUMBSEtTAjK4HxLMBkXpjHv8qnMmjWLyflTKPVO4E15HWVGNjmyFCsaitSxCS9XxB7fCOZYHFlosqmyxfchd7jn3bE+5M5Fwiwq8UHWQeL7yCOPsHjx4lMea8aMGcfdf+WVV9LW1nbK4/bj6e2h09WKp9fMlDnSd/i/mbO+CCeECAeWA/dJKdsH2lpKKaUQYlgqQaSUTwNPg1mIMRxjnixDtXQ/lmvZyTLr9kWU9zzDqzVx6EKhoCGJSdWuIeOqSVk56CHh/jbvA2edJ+Ky9BRKXn+TmkgnWVol38m3kpb2MC/usLFsh5Usy53MjtjEmOT5zJw1D6WrmO1Fl1MvY8DwWeMZEsXlQXfYENLAKjScaislyaMZ2dZKUnAsANGjZhD+6iJaHduJaptI9+x8yrt7iU4I5dNdm1hXto/4lFG8mvU1dAEWeT13Nj/Lg7G/Zjc5jJW7SfPOB049bWyojh/dU+Jxb9IxBChSJ3nK6RWxnCwDF0vP9O9jONm4ceNx97/33nunPbantwdXXS0SiXAJnEkjjvIdHkj/3+9/C2dVgIUQVkzxXSqlfMO3uVEIkSilrPeFGA75ttcCAxVjpG9bLQyyGBiJGbY4Zxiqpftw/AdrGDkF/WA9EoFHM3i/pJLc1NwhHzuwzXv/rPNkBFipdnF+YzLKoT4my4nEjEvnnYMxLH6vBIxg6hTzddRuOkBQVCHPbNjMRm+mr/eE73NOEUinBQte5rCGNFnFS8l34EWlREpu6LQxCjikuHmpy0OdO5dgRafwrW2+kAZIQ0OXEYjCQxh5sUhnCDpwIPoiZim/IssoR1GsOJ3TT+ta5qY6BxWV5KY6wVe+/HnEgM/GhzTAb3/7W1544QXi4uJITk4mNzeXffv2sWjRIpqamggNDeUf//gHY8aMobGxkbvvvpvKykoAlixZwowZMwgPD6ezs5P6+npuvPFG2tvb0TSNJUuWMGvWLNLS0igsLCQmJoY///nPfuOeO+64g/vuu4+qqiquuOIKLrjgAjZu3MiIESN48803CQkJwdPTg0SaWTlC8uor/z7KdzgnJ4cbb7yRDz/8kJ/+9KdERUXx8MMP09fXx6hRo3juuecIDw+nqKiI+++/n87OTmJiYnj++edJTEw842v4RXI2syAE8CywW0r55wG73gJuAX7v+/3mgO33CiFewVyEc/tE+n3gMSFE/9TvUuDBs3Xep8NQcdjhIEFpR8HAwEzVSlDaj/nYpOBYFENgIImMaCJUfIzbHXZUkcG25X+kbt2HJM26hEnX/ZTPCmr4jbTgxcIHEoILangrpW3A7Baq1Egmhu3n7fXb2OhN9hn0mJ3YxkWVMWXULtrt0YwVexktd/OWWIDepqG4ejAcNopGqcwBPiyp4L3eTHQUFCQSA4mCoRuYxdQKGAaW1l40hxULOqP17aiJz7KxbB+zskedUdFEYl8DeW07SOybQL/pzucV8z0bH9JFRUW88sorbN26FU3TmDp1Krm5udx111387W9/Y/To0RQUFHDPPfewZs0afvjDHzJnzhxWrFiBrut0dnYOGm/ZsmVcdtll/PznP0fXdbq7u4863nPPPUdBQQFSSqZPn86cOXNwOp2Ul5fzr3/9i3/84x/ccMMNLF++nJtvvhlbSAjCJZBCIhDccNONPPP880f5DkdHR1NcXExzczMLFizgo48+IiwsjD/84Q/8+c9/5sEHH+QHP/gBb775JrGxsfz73//m5z//+bC6uH0RnM0Z8EzgO8AOIUS/2/LPMIX3VSHE7UA10N869z3MFLQKzDS0WwGklK1CiN8AW3yP+3X/gty5Qn8cdri/Xl4yJZOSkteo84aRZO0i7qDCsns/HnK2FtsZRkpcCluS2ogL34Kr7z8UFb9C7tSX2XggiH+X7kE2HWBWwVaSGsfiLlhLodrAZ8TgFRMxEHgEPCk9TAyrBRHpT3U6WBdPR5LOuk4dA8XfdUMgmBq/lbnOz5ASqt1X8G7zCPqsGpayFtPkTIGouFjq9u5m+75DGIQB5geFKiRS6keZ8VxseZeDlSnkOkuIlAe5c9n5eA0HL5W4WHrH0GGYE1G3dzf/+uNv8QSFYHv3Lb75059/rmGAs/EhvW7dOq699lpCQ0MBuPrqq+nt7WXjxo184xvf8D+ur68PgDVr1vDiiy8Cpk2k3T44FW/atGncdttteL1evv71rzN58uRB+9evX8+1115LWJhZ+LNgwQLWrVvH1VdfTXp6uv/xubm5/ri/LTgEZ9IIPD092EJCsAWHDPlabrzxRsD0Hi4tLWXmzJkAeDwezj//fMrKyti5cyeXXHIJYIYqvuyzXzi7WRDrOXYz24uGeLwEhpyOSCn/CZzTH3VDee+eKWpPJ19PjKNXs2Dpk1SWpGEo6pAr9gUxB/hlTg4aFt7ifH7GI2Tq5bxR8hG/bJ6Jui0YjCzWjRxF3BjBVGM6C0IfJis1BbUlB2lYMBSFsoxEyoMsjEkqoKI2FRAYUqG4KZOySROwFLaa6U++43Z5w5ESKlzp/N+ii9GkghCmVzAIhCGpL6nhtS2PY1WiURPmowsVFYNvjn6dLj2MLEcFNu8s6oMzSYhK4fcfzEIzFMqUNCYmluLRzZmyVzP4cPNutOKDx/ygO3KG38/2zQUwJoJYxyHcbXFs31zwuQrw2fqQPhLDMHA4HKfVYWL27Nl8+umnvPvuuyxcuJD777+f7373uyf13CM9gHt6Dpem24KPLbz99Iu6lJJLLrmEf/1rcJLTjh07GDdu3CCXtq8CgQTCc5S6vbt59y9/o6tKI74mlDEN2ditlmOu2K9pr0bDgiFUNFR2yXEYUmFjayK0af6Qgi5V6jsU3u2xs7bjAjKjqvhJ7hOMydyPlheF4QxGR8ETFoNUzCivoSj0JsRjOIPx5tjNxsOARdHIdpYjJWysmGc2vUTFkIq/c7KhCDqNzeiaRkJvPdfUv83kjlJuaG0naWsDU5q3IdZDQr1K+poy9mypRjNUX76xSrHIQxUgpESRBq7/vMiGV1/mtd/8fFDGybpnn+TlW/8/6p8sQ5aOxb14LduW/9G/34hsZcKk1aSlb2XCpNUYkZ//l6ikrBymX3vDsInv7NmzWblyJT09PXR0dPD2228TGhpKeno6r732GmAK2rZt2wDTk3fJkiWAOYN0u92DxquuriY+Pp4777yTO+64g+Li4kH7Z82axcqVK+nu7qarq4sVK1Ywa9asUz7vY/kOg+k9vGHDBioqKgDo6upi7969ZGdn09TU5Bdgr9fLrl27TvnY5xqBUuRzlO2bC6iPdVInqtjGAa7wTCbGGoGr1wChYA0ZnOgxsqEVi1NDkxILOnEt7ZTWXc683AxW+UIB0ud73h/bLWrLZa7jU0Y5qrjGuYpdTMeQOqqUhITrePOiEC4NnCoxkbtQZQ76yFC0cJWc2iLCRSevHlpA+65IRLfFNys2EIrEkx2B1AQ4VVq0TuIsFjSvRkJfI6ONVPSwBLTuW2jdfYAwrY3q7jwMRSWoUkM4TCMeFIGeFM6kXhuhNd2M6CwltvMgEgbFUf1FK7Zs3KPN2LRi6CR88gqTfBXHI1I9NDWbucVCGIxI9XxO7+TZY+rUqdx4441MmjSJuLg4f7ufpUuX8v3vf59HH30Ur9fLTTfdxKRJk3j88ce56667ePbZZ1FVlSVLlnD++ef7x1u7di1/+tOfsFqthIeH+8MVA4+3cOFCzjvvPMBchJsyZYo/3HCyHMt3GEzv4eeff55vfvOb/tDJo48+SlZWFq+//jo//OEPcbvdaJrGfffdx7hx4071sp1TBPyAzxGOTFF67bln2FVVA0IgJEzxplPSbWe3tY+UPiuTja1864lf+5//7orlrG7cQVNsJLFN7eQQz+VXzad187/4oGQva1Ivor4rGFejzZy+KvD16V0I6gj3bqfbZSPMqdHafQnjOlzIrLf5g/oQGioKBrPlx0Q0e6jvzCCqo571hzLxGoc7GAgkQsJ53lZSR5fzSvZ8NFQs6Pzw0FK+NfX7vP+Pv9HVPBlhS6Z/EU9IibNlB60xE0GoYOiUpZaw0j4VQyooI2zcWn6AuP0JGN6DeDpfQyhiUCbBsnsfwuWdZT6/vzzM0EnNqmf+T8yv0G53MUXFN/vsNq3kTn35jBb0An7A/70E/IC/YgyVohRhtXC44F/waU8dK4Kj0AljS6hkZPzhHMqiahflIgnHoV1ENrSjqiqXL5zPxjUrKXy3HWv8FErTx+BRJGpbDxmte8h2lLMi/Dr0tlQsxaNRDIkNyV8IJzGtkE1CcgEf45YOtomprBUXI2Ik2doBWpuS0QzrgMU46E9Ki48YQVBINDfrz9GpRDDGKCVnxA9IysohqMdOt20kCIXgqHJCY/fS3ZRFKxPAkCDMgpCsuFqs+8egGwpqfSchjgqQ8SiWEcTY4hiz4LJBcdTk/nxeBcypvo5igbwF1/ivkd0+ldypLwe8fwOcUwQE+BxgqBSliedNZ9faj+izhWDp7ebA+AXoh8z0LwOocWQCg0tsEy1j+MF5DuwJKSwpbOPfpakY6ekIBTS3BxkVhOEIZpJ9D0iBJiyItk4wQCLoQ/BxnJUJQaP4nXI9GhaEkOgoIBRAUpqYjgjxYNvSDL7FuP6cYEXAynQbhmZFuG/he23/ISf7B8ybdTlgCmXr+oPYorbinbKbLW0ZZI18g+CS6+htTvev7G04lINmqAhAM1Q2d4/lIt+sNnrSHKZfe8Og6zdrQD5v1Og0YrMvHNJwJ+D9+/myaNEiNmzYMGjbj370o4CN5QACAnwOMFSKUlJWDjct/CltW2twTE5me5CNtcvLfZldkgvHm63bB5bYRnsjcVXbeWhjBR4Dn2gKM72rrhvh6kNxWMnuDiaiYxRvxvdh9GjmopovLNGWHk6D40I02YUhBIqUKFJi9Kc+CIF02NBSw7BUdYEvK3hk2EGc0WFs3ytQDUDAe6FTSeo9yDzfOs2oWRdS+NH/we2w8nzJ99EMFYuic7tjBxEtGSAUpJQ0GWYBSX9wrM3qBI6e1Q7kdPN5h6upaYCjCZizn5iAAJ8DDJWi1FfdztaNnRTZY8jd2MnFN47nietgzc6DXDh+JFdMGwOYJbYLOjcwtS+S8PSZvNDVjVf6eq+b2bUoSERtJwoCBUkU05FILDVNGChIBfQRoeiJoUSlO0goKsBqT8No60M4bFxQv5f9MQkciIsyZ8MIsPQn0AgEBtMStlLpHgtGlDkrllDXpfK7rnh44hluu/cOanbtQALV7ZPRVDNjQjegui2TCYbuS9Y3CO7pzzI2X0Gs2kr+16cNu43kcDY1DRDgdAgI8DAwHDX+R+YRv72tlv+ZGoJXgFXC79aX8q1v5/uFt5+alc8zqnc2rhGhPDjahlGvobSBIs2wwNjgGoJD2ilyjaM/TvuZMCucDGkm8GMAwSrW6GC+kR7LxqfrsUWG4wUUDBzWRq73bKY0Jor31GvMun6n1VdIIREKVESNZaJzKztb0tCl4o8La0i21ps5of0z/Yld4XwSqeDFwIrCvHAbiaMqqClpxNsdRHZfCqVWBwYCFcllY2PIvTzttK7r8djY1onXkOjmxWBjW2dAgAN8rgQE+Aw5mRr/ur272fXJGgDGzbnwhCJdt3c377btxxuWiKEINEPyfv02PN97leDkSBpD4pkwLZcLZ5+Hq7SZGqtgrdIHRR0ovlWxxFAbV1oL6B29j3X156O0GUgpUYCMkE6CguA9VwgGoArJdOturrcmk2efxCs2J/oAt4cGGUlKWgf/URciUUAIdKeNcblVbG0bj261UOyyUxKVy3WjVlFZl8KO7iQwpZq4OLPktX+m37hpD1Fl8JnsZYYI5ZpvX+q3z1z37JO0bu7iitgG6mUkicJNlJh00u/HsQoxhmI4m5oGCHA6BAT4DDlRjX/d3t28+qufoWteAHat/ZAbHv7dUSLcbwzuSLdRvrUYZ1QiauJCwGwiGX1gG23ddvQdW5FobNnyPvAz2jJH8mqzhtY94Gu7hDijnJ7R+3i5+Hp/SfCUpB1MC9pDXs7tzJo1C+0vf6OkrZfzJrzLaGclRrfK5vfqcATvQFXmgqGiKjqO4FZWN59vegM7zewMBUmEswtNCcZW2GpmMShQm5WMVQlFDgghlFpH+19nv3NbZfULxOg6+1SVQwPaDM26fREHCh5gppZPo9pOvBZP+ZoN8LXLTvhebFv+R9yL1yIjzFLrbXBcET5WU9MAAT4vAgJ8hpyoxt8UaK//vq7rR4n0wO4XzdsK0DweRjRUc+F7K6iwZ5LZVs6IxjpCR2g0jo5kl2UKcXtrCN5ShC3vEvT3dwOKb74qsSgSS5bKurYLBpnq9IWEcl7aBuLjzZ4SP/zx3Xz0xo/AXokQUOlOYXt1E3Ej+vhJ0hOUuUbTZQnjw7K5GJ0Ca3Ur3jwnwmHjFvkPUow61rXmgyH9xwhvt1EXoUK3r/BDgWZr0MBLQlVVFbVhkdTaoxnhbjnKuS0pKJUE6SRJj8aQOl3aydX8V35SSsOAzscDCzGOxZFNTf+bGeh6Ntz09fVx1VVX0dzczIMPPuj3fhhOVq5cSVZWFmPHjh32sc8WAQE+Q05U428KtNUvwqqqHi3SA7pfCEsK9BXQEJzAels+eq9KXXAsoeNH4cgv4qWOu6HVC/nTWeTtIK78NSzKGHRDRSCJi2ymdsQoNiSMRG3rQVXa/EKYF1WCEGCz7aeo2sWbH2wguSOC1Akqle4U/m/xIjTDCmIC4xP3Mz5tJ69XzUfIbjMBwpBMbCxlRm8JI7qb2a1Ng5hgUDqRBlgVOC99LB/srkLPijCbUjksXGIfXOzTHj+CtyeGoiuCYkMyP36wuU5YdjBGDSANDARh2cGHr1VNDVVVVYNaHfVjTbkQo/xw52NryrnV+bivuv2kW0WdKlKa8XhFOTfdBUpKSgBOyaPiVL2BV65cyfz58wMC/N/G8Yx4krJyuOHhx44bA04eYAxuUeMYdd50dovx6IfMPz4DhXcc6RgdY7EUuvwhhYKRGgu1Lfxk6mo21k9jQ9106tvjEJ1dEBmC7ghmSu4+NBfkOkuYa18HusrG7VaWrF3BVXVv0SZ1lKbp7IybiGbYfK5kUNIwnuK2SUirwKoIpCERSCw1Nhrq0jDUWMqTR6FFWWFaDGpLH1MF/HldA7oIxdbaRmaih/EHJfnJhzDN8Uz2BYVhqO3msVTz/kDOu/dWXnz232ztDWdycCffvd3MG62pqTmu6XzegmuoWbwZwzh+ytoXQV91O83P7PA3S425Y8IZi3BVVRWXXXYZ06dPp6ioiPPOO48dO3bQ09PD9ddfz69+9SvAnNnecsstvP3223i9Xl577TXGjBlDS0sL3/zmN6mtreX8889nYFXssXx/L7/8cvLz89m4cSPTpk3j1ltv5eGHH+bQoUMsXbrUX6Y8kEOHDnHzzTfT1NTE5MmTWb58OVVVVfzkJz9B0zSmTZvGkiVLCAoKIi0t7aS8gR944AHeeustLBYLl156KQsWLOCtt97ik08+4dFHH2X58uWMGjXqjK7v50FAgD8HTuSUNusoY/Bf8PErH8OhLvqrEyQKtGqDQgrhDV04s3IIi1xOmSsTXSqYnrqgtvQiIkO50rGKLOdeyowsXun5LjOqkmmp8xDfXY8qdRQkrZVthNmjUFSBppuxXLW221+yrGVHoLZ7sNR1sctwsttwcBllZHr3UUImml1FsQfRu3Enuojz9ZcziC1zM0pLoaZ5x6DXO8MRjk0R/vSvIxe/Ct1dPJQ5Bq8heUMRjHV3kWcPO6HpfEKGnWvP0c7HfZVupGaYxuSaYc6Eh2EWXF5ezgsvvEB+fj6tra1ERUWh6zoXXXQR27dvZ+LEiQDExMRQXFzMU089xeLFi3nmmWf41a9+xQUXXMBDDz3Eu+++y7PPPgsc3/e3oqKC1157jX/+859MmzaNZcuWsX79et566y0ee+wxVq5cedQ5xsXF8cwzz7B48WLeeecdent7mTt3LqtXryYrK4vvfve7LFmyhPvuuw84sTfwokWLWLFiBXv27EEIQVtbGw6Hg6uvvpr58+dz/fXXn/F1/bwICPA5wsBCgkJ3F90x7ShCoEvVVNwRNoQUflMdi4CvJ3bT1RjCpp2X0xjjQBESXRpYhEFeZAXpFXUYGQrPKnfxqTIPPdTC+zk6Pwspp7HXit6mgtTRhUpO9ijY7UGLDUb06qi13X6hD27sID1kP/tkimmgjqQjPJmrHe+QaWxitzKWMUYpTfY0yjsuR5cGqjRI0UKGbPWTZw/j0YR4Pqjdw9ywCkajM7DN0LHSwwaw65vzAAAgAElEQVSazitS+FsdDSQhw35OCW8/QRl2hEXxz4CDhukcU1NTyc/PB+DVV1/l6aefRtM06uvrKS0t9QvwggULANOr9403zOY0n376qf/2VVddhdNphoJO5Ps7YYIZQhs3bhwXXXQRQggmTJhw0qY8ZWVlpKenk5WVBcAtt9zCk08+6RfgE3kD2+12goODuf3225k/fz7z588//Qv4BRMQ4HOM/uIAj5qCZVoXU1o/I8LZw0z7Zzj3X8H6+Kkorb1cHOLG0t3AR3sO8Wb0xejN/V0qzIW4qJomLM6x/E65As1XEYcwW86Xpm5j8QVX8p+35mM7uIu0KVM42N5GdGcfCXEHaUxKoqE+0u+eprVCuZKORWgY0kABrhwbT6ScxGieIdPYi5QKvZ505idpVASFMr63hgndNUOaxxdVu1j61nJ+NPl/sbRpFBbZyMs9bI5zrPSw2M4w0mwJFFnc5Gp2Yju/PItnQamRxNwxYdhjwP0iuX//fhYvXsyWLVtwOp0sXLiQ3t7ew8f3+fWqqoqmaad9vIG+v4qi+O8rinJG4w7kRN7AAJs3b2b16tW8/vrrPPHEE6xZs2ZYjv15ExDgs8DxForg+IUbG9s68RgSA4HmCCHV3sg1YgWGATVdecS5ahkTU8H5l85mxR/epy5qwuHqNF/MwJAQHnGA8sh0U3z7W85LAws6OWIHIR2hOLZvo88SRFllBSgK10c3Mzntfd4R17A8+2rU6m5Et+6fCY+M0Ug51ER0gsrKPe2MPFjFxPDp1Kao7O9OQlOjWD0hHV0IdogEcjLiGTVm9FGvf1NlCxeGbscqNLNixPDStP8T7JNNAc6zh7EkysKamjouTE4iqfEABWt20Kgo/LorFs1I4F1FJ8lSwUWcuO9dP/2pfme7/9uxCEqNHPbFt37a29sJCwvDbrfT2NjIqlWrmDt37nGfM3v2bJYtW8YvfvELVq1ahcvlAkzf34ULF/LAAw8gpWTFihW89NJLw3au2dnZVFVVUVFRQWZmJi+99BJz5sw56nH5+fksWrTI/7iuri5qa2tJSkqiu7ubK6+8kpkzZ5KRkQEc32f4XCUgwMPMiRaKTlS4Mbq9GVXTQFFQpEG23I0hBJ8enMHS+mlIBJb6sdxU+iKuazNx7vGgthiY0VxfvFhR6IyKZFNUBv1GD4rUmSNXM4tPGCUrKN2Zx760sdTEpxDf10NCh4soRz2KkNjdLVjKOszFPnyeDApkadHEWKJ4aUIkugLqpFm0f7aCj6unoxkWFEWitXkwnEEYus47pWW0vvo8Nz40+DXmqK20NgahZykIw0BKQVdNPEw+fI32/vGXJGoaZYpKGRLDMNgzJgvNmO0vYd7UVHd0a5VjMDDVb6iOIl92Jk2axJQpUxgzZgzJycn+r+3H4+GHH+ab3/wm48aNY8aMGaSkpADD5/t7LIKDg3nuuef4xje+4V+Eu/vuu4963LG8gSMiIrjmmmvo7e1FSsmf/2y2nLzpppu48847+etf/8rrr78eWIT7b+REC0UnKtxQ1+3gqbYkSqJtpLe6aLalsyUilqX7r8XwzXQ1w8ILLd9Cy4jEMqqHzL5y2hQ7jSNizdQvp8ou+1R0LKafsKuP8JomvL0GLmcc+xozWB2bzoa5E01BR+fWupfo6wzGMCx0tQahGIbPRQKM6CCMURG0VvZSFqZjtGvIqCB0YHvMRLw1pjWlbghUVy84bCi6TnJdJZrXy6YNm1kw4DWGtVSxSkmkqvEn5LATR6OONStu6Gvkt+SUxLc1YgnR0Q2wKDqzsk/+P9jAVD9DcFRHkS8jaWlp7Ny503//+eefH/JxA4UzLy+PtWvXAuZi1wcffDDkc+6//37uv//+kz7ekfuOZO7cuYNm5BdddJE/Ne1Y5wpw4YUXsmXLlqMet3nz5qO2zZw5k9LS0mOew7lIQICHmbS0NFRV9c+A09LSANMQ3OUqIDoz3l+4UZ+YxocpOajuLkSbh02VLYjOaNLcGtZ2DzEyiBzlMl7W9mDIATFeAXpUMNKtIQrdVBqxCCEg3IruDEJgoHUHITwe1Lpu1Npu+qSgkAkUto0neYxCxchEpNkeAl1K3ElBzDSK2Fg2H9UajRBmDA5FYGREoBiwtbkTHbAW9qKPiUSMDCU4RDncbUOB2YkW9M4u4la/QmLjQXShUhucNOgatWaO41/hGeiqwkZjKl9v2sQ3Jmf79w8sblEUFXwz4JjmLu75jo31zTpXZCUye1z+Sb8vA1P9hloYDBDgiyAgwMNIUbWLTZV9TL9sAcG9Lf4YsNtdTOGAbgxX/fRRNuy38HhkClq7zjMf78JW2IKmGygCEGYGgFVIHknYRlK3E7N9nzkbPC+lmM8clyMqe31hAmG28HF5wBmERKG2LxZbUbM/jOA3TZeCSm+EX3z7y4rHip1YhY5h97As8RJkrIba0sf4rgOENQo8NVa2+3wmhJTYyjr48SgnnzW1UZkYixSgJITwtaomRl81ix9/mku9M57G8JH8ZcbgRgG79XAMxWP2jROCmPPmDq6EO6K4BcxZcWvmOL7fquFNlBR1CSb70tNOhqNT/RYNeM9ayM+IPq1uywEG89xzz/H4448P2jZz5syANeUxCAjwMDHQGD0/rJhv5xygz3EFJCezc/9adN2Dqkh03UujXkVpygI8rS6kECgtvXh1AynNIghTZgVeJI8rKfSqNt9RTOvHEdZ6HpC/4qPQiykRWebsWBFIq4K6vxMjygYuDTEghjuwFs2I9q1k+77eX6G/y2hRgZQKW3qmIvZ0mCGFxBCsQuX8zSU0iwx2OiJ8em4WZlTs72BzayIWVzeKEPzIpRDWVE7uPVfzlx98/ZjCluvSsUqJV0qsEi70hB51PY/MnU7KyuGv1Y14muowhEDq+im7lw2VjdH/ntksCkvvyA+I8Bly6623BgzXT4GAAA8T/cboucGFXDB9M+8rY6luMBPb97Rm0ts2igpXOtmOfYSsLSU8qAvrzEvRpIqIUrCqoOv9/mMaulQxFIWGumi/egokqqKT7Sgn3ThAT306EfEGa8KnoVsVrHvc/h5B3jF25OGIxWEBFiDavUi7DXyG64m1uRwihA96s9hTZUHFtKtUa7uJpArcm4hhM3O0bD6JmWV+jcegbE8ZOtFInyBvqy/EEdTFRy+8Q/rYTBbNG2yd2c/YpFoePPBHPpUXIFDwxMwAMk54jUPXvY8SPwGpSBRDELrufUg9ubbpx3vPDAlezWBTZUtAgAN8rgQEeJjIz4hmpPcQE1Kb+b36EBoWLKpGyP73aS+L5cnmRehCoAI3tAczIet9HjR+xW5lLDmRpaTMv46164KIt35GRPIeVrquYUfPhMMFEYAeFYSR6cCw2xASQkNdnJdZxWp1NsquLr/YIiVKu3dQ14qBTmm2PW1oEQLDEYJFQlmb5PnwfKRXR6W7vzMQSElfr6X/6Izv3E1kKByKiyfbWU5nZxTlrummEZrUSZIuqiOD2F9ZyGeVxcwoySBnbu5RqXZbm9YgFY31Yi6asLCuWfLGSYQTQjdV8J3IZKrig0hr6CO0owJuPrP3zGZR8GoGVotCfkb06Q8WIMBpEBDgYcJdXc7VdW+xP+sqNCwYQkWTsKtzNIdqHGih5pd3DaixGoxqyibTeIfRlCMNQfMb28npnY07VZAeWcPVjlXscOdCfbd/gUsbHYm0W1nHHDJlOTGxB1in5GOgmrFjHxIBHh1CBoYupP8W0iCvdh0dDXE09jkpaLKhNHT4Wg7hD00oGIzsqgYEiqoiBGQFl3JV/koURccwVGzvN7C/fQQjeutIjUmlPbKRSEcD7W0J7N8DO9e9cVSqXXmNnd2xE9CEeZ280jipcELoaCcpO/sY2ayhSJ3Q8Wc2W81NdbL0jvxADDjAF0ZAgIeBomoXf//3B0zTvcSVVWPNmIJXMbAagqSyDrbHOpHd+L0VOhMFvVWj2L/+ewTH76K3IQevexIhjlr2MpqNjROZaCvkO86X+DD3YjpcwbRHxZphA2AtFzNTrsN+KJ6xjt1YhIaeFAS13T7xFKjNfeheXxDYZ4zej0Bha+24Qa+hX3f1EaEkupvJ6DL4mgxG69MInziZGd/4FgD7Kv4Xj7oHAEWVhKT2MuGzPYRgISTmECkTP/SLc11T3pCpdvY6L+2N8VjUNgyhIBKDT8oMvXbi19i395+M6jaoCFUwJt52um+Zn9xU55dWeKuqqpg/f/5x07/OlCuvvJJly5bhcDjO2jH+mwkI8DCwqbKFdG8TApXwgyXctkqhLnUCDW6F8nALZS79cGw2207ByFAyuxsIFx3IQyMRShd2WyeHQuJ4e9J4dEWw0ZiKcHvA5cWICkI6zJgtgCFV9ihjyejpIUR+jQvkWtx2O9uTRqPU9vrFVG3r9yEeKL79bTSF3+bH/68AI9JC/dh0bi6qZnaLnXcix9KQNY30oARyU52Exd9BccknGIYXiYVPumZQkZKOKuDXKa+hCB0hQBE6qmM/am3i0fab1ZL1IdkoosfsU3ewE5HngRPMgPMzovnfqFl4I82QwdIvWcjgRBWS5yLvvffeF30KX2kCAjwM5GdEY/RJPL774QeLyDpYRIawsPaSH0NLp18UhSaRbV62RPYwN6iOCI9KZ3sslqA2aqKy0YWCFALD7cFW6AIJKp1oqWGm4a7XQO3w4Ik3MEY386jIRcf0ApYjvNjq+/zWkYOX38QR9w5Hhs1G92aKmrWsAy3CQk2Cws5Wg/8/MhNvkYt/btvEv+7MJzd1KlOnvITLVUBj31gOdnWjClMQE5IvxuhbD0ikIUhIupCLFtx8VAy4JTnJ511hYiBOagHsyxwyOFGF5OmiaRrf/va3KS4uZty4cbz44ossXryYt99+m56eHmbMmMHf//53hBBs2bKF22+/HUVRuOSSS1i1ahU7d+6ku7ubhQsXsnPnTrKzs6mrq+PJJ58kLy/Pb9Le2dnJFVdcwQUXXMDGjRsZMWIEb775JiEhIcccN8CJOTfdm79k5KY6CY8T+JNufVikl2z3x2ahApi/LQJrUTPVzTaW1l9GaOpeIiObSTVCuLZLxyZBNSSWuu7Di2qAtboTa0UH1upOlFYP7++exyeuaRiKglQUDATCbkPLjhgw4ZUIdKLDWo8658NpaQaqkPTPizEkamsvU41y1lpNXwoJeDSdN4oPAmC3TyUt7ftMz57D0jvyuf/SbJbekc9Fc24mPfGPBHkuIz3xj1zy3aP74wHcdMNVWIThC5dIrOrJL4DlpjpZNC/zSyW+MHSF5HBQVlbGPffcw+7du4mMjOSpp57i3nvvZcuWLezcuZOenh7eeecdwEwR+/vf/87WrVsHGZ0/9dRTOJ1OSktL+c1vfkNRUdGQxyovL2fRokXs2rULh8PB8uXLjztugBMTEOBh4rI770YZtBImEVLhipZVfGfq64zN3MdlU9eT565FMSQSBa9h4dW91+CwOInVbWje/ZxXXExq4R6cTU0DFs788ugPH0gE+xpTUdBRpIYFnZnujSjtHn88QSIwhIXkmA4sQvc92/xRhCQ/tYQFo97lkpTVKMIADFRF50rLu+zdn8x+pQar4kVBx6p4UbsH+/oOVcSQOWEBs658iswJCwY9tq+6nfaPa+irbic31cmvr52EqgwMhXy16a+QFEIMqpA8Uwb6Ptx8882sX7+ejz/+mOnTpzNhwgTWrFnDrl27aGtro6Ojg/PPPx+Ab33rW/4x1q9fz003mW2qxo8f77ewPJL09HQmTzYNO3Jzc6mqqjruuAFOTCAEMUwkZeWQnXceB3cdoE6GUhEyjmQ9jCZLA7Ptf2Ou81MqXOls7kxCESmmzy+C/e1p/LYjhe8mVvKiOw2te6i3pD+ccPgegFt1QE0v0d4W8h0F9IhQlLreQc+QEor0HMZFl9LrEVS3m56+QuqMsBwkO7qcxYX3IqWCKiSXJH/MmrJ5eKXAIkK4aczrdHnDyXJUkBg+ETD/g51KEcNQ3SBc3R7/69AN+ZXPwU1OTuaWW24Z9hiwEOKo+/fccw+FhYUkJyfzyCOPDLKlPBMGWlGqqkpPT8+wjPvfTGAGPEyse/ZJqsqmURV8PS/HXcKGiCRec9jZ1xtPZUUen9TM4E9FP+Tj9jSfrwP0531pUrBJ0dB8onz4p7+KTZimZkccU23qQ93dQXuFjQ+KZ9FQFzUobNG/sKbU9rCnOY2ajpEowkBgIJBkOSsoaxmNZlhMo3UpqG3IxSvNDhy6VOj0RHBF2kdk2A+Qk3Gp/9hDFTEci6G6QfTn4KqC/5oc3OTkZGbNmjWsC3AHDhzgs88+A2DZsmVccMEFgNkBo7Ozk9dffx0Ah8NBREQEBQUFALzyyiv+MWbOnMmrr74KQGlpKTt2DP6mczyON26AExOYAQ8T+4t7kMJCjVVDB6QAXUKVFgGag6Vll2HIw769A9u2CyAlooa9rkw0OXCm67utgJYcilrTjTAY9Dz/bwMau2JRhT7gOAwqgdOkhWznAaI93czO+Q+Zzv3UkoKhmG2MpKIwytnNnkNh6Aaoik6kGMGbrv/hvKZQ5mTNAGDVjl1scx0gOqyT1q7wEwroUN0gclMjv7QLaucS2dnZPPnkk9x2222MHTuW73//+7hcLsaPH09CQgLTpk3zP/bZZ5/lzjvvRFEU5syZg91uduW45557uOWWWxg7dixjxoxh3Lhx/n0nw7HGDXBixMBGfF8V8vLyZGFh4ed6zH/e9QQ9Ioet0QqrjW6klKgIvtHmpi2tnFXuKYDp7AUDxBVfibEwuCRlNZtdebS0Rx0lsgNTGAbu68e8bwy4rwwuQfaRluTiO2ErSEvbihCwUi7gDfd14PKCw8qClgrGO1+irC2DUKeXlxz34BUCq4QXVTvbyt7jjxnnoasqqmHwbXcnN0ybfEIB3VRwkI07G5kxPp786SNP9fKec+Tl5fF5/42dKZ2dnYSHm/nWv//976mvr+fxxx9H13W8Xi/BwcHs27ePiy++mLKyMmw22wlGPP64X1WGeu+FEEVSyrxjPOWYBGbAw4Sls4WDaRZWzYvEaA/F0trLheVNxLS9SpxD4z/uycij5LAfgS4VPjhwEcHR3n4PdWCAgB7p6cDR4jqglM2fhiaOmG3b2iRubzyGoSKEwRhZihJ5E5rdhqprhH+6H0vYdUyL28sa5uJ1CgxFoBmSFQfLqe+IQVdVpKKiAy2u/ZT96WO6j9NloqjaxcJ3duDRDJ6uOsTShLDAjPcL4N133+V3v/sdmqaRmprq9/Pt7u5m3rx5eL1epJQ89dRTJy2+xxs3wIkJCPAwocfUUJmgoSsgo4LwOqzUGA20R80g1l3L4XC7GPDvwDIIgSEVupuDjhzajzzi99F5vYLB+b+HRRjAiuCC1ki8HRPZIQV2RwMedyLfbX6PvfZgQhvc1HsTUNpSSGrJICZKwZIh0RFYFEHIbitpupUNuoEOqLpOzN4wXN5Zx+0yETC9OTe48cYb/Q0vBxIREXFGs/ljjRvgxAQE+AxZ+fhfcJU20x6bA10VWCrHoEUFgcPGzsxRSJGJdV8HSlsn+KXQRAwKRByeqZr7BosrHD3jPXbwyEAVBlIq/V3iyFRaGK22Mt6TSUN7Jgd649jS4CW1z8rIrna0Q20sj8lHCwIVDze5ipieEMp1eXexsa0T+d46LC12EE6+vfoAVTFeshtriO+YfMIuEwHTmwABhiYgwGfAysf/QnWxgaG1UR3sYIuWhOruQKnsxDMtBsNXPmxoA6UVDBVkhBWlR0f0GUfHewdwLCG2hmloXWbWhGRwTDgx8hALs5fh3T+OksbJdEZ76YkNR2+zUN/jokGx8UaoBQ0bxeGSB7I1XPuC0YV5FB2VZVGTSZqYzdftYeTZw1ix+x0ajJswBDi7IzDqGskamUCH+8RdJnJTnfzl6nTW7DzIheNHBma/AQL4CKShnQF1m/ej9X6MoVVToDgx+mOwEtQqszurkDpKq1mk7JdgRxDa9Di0UREAR8yKGbRNDrENwCODfMUTh/PO+lPW6joSqJapBHVZcFptbMjLZUt6Dm9Pmkmj7RB9GS3oiooUAh2FnnFz+d59N2K1HK5isqpi0Ew1Y85Yxu38K6KrkOdCvSwPiuI3TVFEjGnCaVvH+Px9x4wB19TU8Pa69ykzGnl73fvU1NSc9DUOEOCrzFmbAQsh/gnMBw5JKcf7tkUB/wbSgCrgBimlS5jZ5I8DVwLdwEIpZbHvObcAv/AN+6iU8oWzdc6nildvA6A+KJ56a/wgtVTavFi2NxPS2onXY/UviwFIq0A50Ina2IMeZUP1CfSgrhWhKkq3ftR2/2y4W0cXKkasDWlToc9Abe7zF198WnMxo1x7MGaFm+XKQsFQIPZ7tzATG28/s2lQSCA31cm/7sznjeKDSOC6qYNnqpOu+ymfNLezZ1c0Xsyiay9wyDGW+5/4znGv0/PvreLNiTMxFIUiwyDxvVX88nt3nfL1DhDgq8bZDEE8DzwBvDhg2wPAainl74UQD/ju/x/gCmC072c6sASY7hPsh4E8TO0pEkK8JaV0ncXzPiENlW42LF2FYhkN3gL2hGfha+aGv1jYo6E06HixAoNnsaLNg7VhcHVSf6jBCFXRxjtRWvtQKjoGhSD0hCCER6K0evxCqzR7/DaX/uwJBcJsIYRoMLanHVuMiteQWFWVGY5wcu1hQ+bgnsiaMThuInFbSrCGjMArwSpgxvgTN7cs10IxhO9DQJj3AwQIcBZDEFLKT4EjXWCuAfpnsC8AXx+w/UVpsglwCCESgcuAD6WUrT7R/RC4/Gyd88nQUOlm+R8LaDgYC0Fp9NjncMg+epBXg3l7cNaDuc9E6TWO2HM411cb72S8fRdp9np/dLff/cHa3IMeH3LY3MenzqaJjunlq42OxJMXw44xqeyInIf3kzju372Ob/U0sSTK4jc9Px1Tm+RxE4joK+M7Leu5uq+ev86MO6mc3tl2A1VKhDR/z7YbJ3zOVxG3u5iqqiW43cWf2zGff/557r333lN+3iOPPMLixYvPwhmZPPbYY2dt7C8Tn/ciXLyUst53uwHonz6NAAYGBg/6th1r+1EIIe4C7gJISUkZxlMeTOEbb4JMwtDraaOUl50X0DcwZ7c/Huu7MzAPd9D5Mji04N8INLc5UdweQiN66eoI8e+UmsBa5kaPDoIglaBwL9per79jhpEYguEMMn2DDYP9sX0kNRzC+/H7JCofsNdiYdIvh3YoOxmO7FZ8suNMmJbKgwceokzmkC13M2HaT0/r+F9m3O5iiku+g2F4UBQbU6e8hN0+9Ys+rS+Mxx57jJ/97Gdf9Gl84Xxhi3BSyv7F++Ea72kpZZ6UMi82Nna4hj0K74E1CEPH8NbQFJaBVwwwdoy0oieF4h1jRx8Zih6moljM2d7AxbRjZTtggGWPm9ZCG80V4XR1hhzeN+AxalMfal03mR4Xeq4TPTMcPdfJ1IPFKIaBMAxUXSP5YCVSO4Dpz2v4u1OcCUlZOfy/9s48LIor3f+fU90NsgkoCgq4EPcVRHHBDeNEEycuibmGmJk4Zpv8NDNZJjN6s5lk7lxjZu4vdybOzWTPNRqNYzIaM06MSYz7AhGMu0JQRERFBFGW7q5z/6ii6WZRUbaW83mefrrqVPWpt5ri5fCe93zfodP+rU5OPDNnM93FYaaIz+gujpCZs5lTRw6y87NPOHXk4A3Z4y0UFOxE18sBHV23U1Cws176nTp1KvHx8fTt25e33noLMErD9+jRg4SEBLZu3eo69/PPP2fo0KHExcUxfvx48vJqThusID09neHDh9O9e3fefvttAKSUPPPMM/Tr14/+/fuzYsWKK7bn5uYyevRoYmNj6devH5s3b2bevHmUlJQQGxvLzJkz6+V78FYaewScJ4ToIKXMNUMMZ8z2HMBdoSTKbMsBxlZp39gIdtZKzJg+tP3jf3OyfR+KOpxGiHYIaQiiUwQU2bHkVJ6vm3/janK6eisNrVT3OKYV2SvPN7121VVxFbq91h+t3H/8K4oSRzIyyMqZT9fTI+MHsjt2pVPucXpf9qE8yg/nCR+k04HFaq1WnaIxiIkcxfkTH4HuxCkttLd0YeUrz+IoLwcEvROGcsfTz121H28mNHQomuaDrtvRNBuhoUPrpd/33nuPNm3aUFJSwpAhQ5g0aRIvvvgiqampBAcHk5SURFxcHAAjR45kx44dCCF45513WLRoEX/6059q7Xvv3r3s2LGDS5cuERcXx6RJk9i+fTtpaWmkp6dz7tw5hgwZwujRo9m2bVuN7cuWLWPChAk8++yzOJ1OLl++zKhRo3jjjTdIS0url+/Am2lsB7wGeABYaL6vdmufK4RYjjEJV2g66S+BPwghKgKVtwHzG9lmDwbe/VvSgUubv8I3OgJxssIxVne0njq+nuvTEKCVVY6O8fhMpYC7IzoAy9lSxGVnlWOC9LgO9Ny8gZ4fL2bs8/8BL/4nu1av4lLBaVonjuOZH/wpd+hERXTnib4awxIT6hx+OHXkYJ1DDlUZ2nMMO/kbmTmbiYkcBQfyTOdrcCzdUJOrLY3tZiA4uLKSSGjo0HoLP/z5z3/ms88+A4x0vyVLljB27Fgq/gucMWMGR44cAeDkyZPMmDGD3NxcysvL6dq16xX7njJlCn5+fvj5+ZGUlMSuXbvYsmULycnJWCwWwsPDGTNmDLt37661fciQIcyePRu73c7UqVNdesIKg4ZMQ/sYY/QaJoQ4iZHNsBD4RAjxIHAc+Dfz9H9ipKAdw0hD+wWAlPK8EOIVYLd53stSyurlHRqZgXf/lqILAWSezcYpNWqK5NQUbnBlM4T5gjBCCdUdtrFhCbfRzv88p7NwFdqoOObs6I8zyh9aW8nu2JWoszkuJ3l87/c4HQ5OZ/1IaPhPyfWNIMfWntwuPenYo1ud7vPUkYOsfOVZnL7RvIoAACAASURBVA5j9FxR3Xjzu4vJ3pNH9BX0H6oytOcYhvYcY/QrDrp9QxaEtRPZe47VyTZvJDh4UL3GfTdu3MiGDRvYvn07/v7+jB07ll69enHgwIEaz3/88cd56qmnmDx5Mhs3bmTBggVX7L8mreG6Mnr0aDZt2sQXX3zBrFmzeOqpp/j5z39e535uVhoyCyJZStlBSmmTUkZJKd+VUuZLKW+VUnaXUo6vcKZm9sMcKeUtUsr+UsoUt37ek1J2M1/vN5S97mx+dzHL5r7A5ncXe7SXHS9i72ufsGb+fHKKjhFd7oPFrCThrupQ03ulwi+IEgeWs2Uex13bAuy9gkngCGHnC0Gv/FxAkJ0+4eeRvVsjW/tg0XU65WYhhMAvKIjs/T/gdDiQug66k07luVgERNrP0CFr61XjrVXv272/ivjx5ncXs2/HLRTYR7Fvxy3VvqNroWOP3vROGIrNdyg+AXdhtbSvdRWdonYKCwsJDQ3F39+fQ4cOsWPHDkpKSvjuu+/Iz8/HbrezcuVKj/MjI4057A8/vHo6/erVqyktLSU/P5+NGzcyZMgQRo0axYoVK3A6nZw9e5ZNmzaRkJBQa/vx48cJDw/n4Ycf5qGHHuL7740MEJvNht1uv4oFNz9qJVwVanMwZceLyHtzD772Yk7H72FXYAnZYRfNEbC7tGT1lWtV47/aJWe1cyvenZH+6NH+OBwO/KyhHvXkOgSXMuz8Ue76chlhR3L5zcU8Is/koDudfP3B2/gFBWGxWhGahtVm5f4xvXncN52puavJWr+Kla88W6sTrum+o/v2d/VXET/OSs1F1yym/oOFrNTcGvu7Gnc8/RxxY0Jp63/siqvoFLUzceJEHA4HvXv3Zt68eQwbNowOHTqwYMEChg8fTmJiIr17V4aNFixYwD333EN8fDxhYWFX7X/AgAEkJSUxbNgwnn/+eTp27Mi0adMYMGAAAwcOZNy4cSxatIiIiIha2zdu3MjAgQOJi4tjxYoV/PrXvwbgkUceYcCAAS1+Ek7pAVdh2dwXKLCPAmEB3Umoz2bue+Nl9r72CSHnItjecRPP5fXFrluonutrRHprF52srvng4Zw1KB8chhZs5b7NX1LQKpAvegxFO1+G3saX8Sf20D/tWyylJXSdcDcAGf/6OxoSHcEtE6czLDGB7P0/4BcUxLcfvo3DbnfN4glNIzKmG87CwGrhg9ruu2oM+J3/WIg9KxZdWNCkE1uXNB56dt51fdfejDfqASvqh/rUA1Yj4CpEx4Wj6U7QnS6Bmc3vLuZAZgg6cOh8Nxy6BUNcHSorXFRse45qqy22oLq+g2v029EfGWKjV+4JkvRQHJ2jkCE2nLe0RgbbuGSBYJ/2BEcO4fzlINIcbY3lyAicwkJOq46uNLGSixdxOhyVKRTCyEU+e6p/jeGDmu4bqqedxd3/KJl+mxCOr8n020Tc/Y/e+JeuULRQlBqaSUrhJbZdKGbE9Fn04wOPSaZlc1+gQO/BJ2E5HAl1YskQOKRO5WjXc3KiarWKCmqakHNtayA7+mKTTmad8ufWB2eSnbaRr3TdVBvTGaD5cqZDO3ScaGf20eWclaMhnTlGGKcDohjVKoLU4wXEdw51hQ+cDgeaptF37E/IScngElE1ykcao+GrT67Fdw6Fp59hR2Y+96pSQl7P+++/X616RWJiIosX1z22r6g7KgSB4XzvSjmEHYENyaeDe7mW7IIRH/36TGveG9oDB1as2RexHLzoWnZck1rZ1eLAFfvOSD8GtUqlf+gBikLa0f6HQDql59IxsR8nDx0mXbNyPLo7nbOP0k8PIyewHClASIgu7ULZ+Qgckan8xR6HQ5ceFYqrhg8q4ry6sCDQCXCkEJMYouKv14EKQbRcVAiinlm24SvsaEhNw47Gsg1feRwf9eAcTkWfw4EVXViQF50uhYaa8MjXBXM5soGzsz+aplceK3dS7NOaLbnDOJYeSUlOFudKstn79b84f/JHok8cYeTWL4g+cQRZGoSGhpCgoSFLQ9CFhcDcyzh0Wa1CcdXwwagH59BvWAaBjl0godiWcN2ZDAqF4sZRDhgI3ZOJRQehSyw6hKVneRw/nVlIlzwN64USrBmFiFKnx/ErVapwjwo7o/xx9AwhuufZStnKs+VkHIwiKyeKA2e7scI2hYIOwWbs1rOn0rJQOhb0o8vlUKIK+lJaEoAmnYT2CbvmEu+jHpyDzb8MqWmuTIbaKlkoFIqGpcXHgFOPF1BUdI6ff3OeH8N96ZpXxqhuhoTk+4//nvOWtlgc7SnWe2M7V4BTVgYTPAsKVccj7CCMFDOA3KJwhKyMFHtMxEkLp4MjCD1dVDmBZp6ll5+gXAyix0ArcJrsPelmvPZJoo8XXHOJ9+i4cAp3XL2ShUKhaFhavAPekZnPqoBEZllXMMWZhz0iBv/xLzD7qbdpFexHe0seyLNkFceaztezkhuA9DGqYAi7rFHzQVrA0SMYGeyDuFCOPaf6Mo2KLYvQGRAUQu9bJ7L36y+RUncpqvnZztC7lpzZq2n5unOtE24KhaJhafEOeFhMWx70WU1C4kbQnGReLOe+t7dS7hOBxRHObdph2otL+GiXEARVE5eUgCiXHvtVvbBwgu1wIdZAO7fknOEoAaYT1/EJclAc1RbtohOQ9PE5QaTej588PA1LzyEs/WgVui7JCOnN/398ar1lHSin2/i4Mm1CAj0meZsLXbp0ISUlhbCwMEaMGMG2bduuq58PPviA2267jY4dO17TteqbtLQ0Tp06xR133AHAmjVrOHDgAPPmNb989RbvgC9vWMYtIQFkFnXi8IVu5JeEUu7QkVjQcVJgs+BvdbCl3B8pqy+0cM920FvbCGxfSrnNF8dBh2cGhA4Rp36k6FIAFYXipaZxKSoMYZc4O/hhCbYSk36RC+ExAIwbnUBw5+7syMxnrhlaqA9xHEXjk1J4ielpx4zKJJrg77HdGsUJOxwOrNa6/5pfr/MFwwH369fvig64IUlLSyMlJcXlgCdPnszkyZObxJar0SIdsLuYTPaePHKCbuGts6Nx6BYsQkcTulHuR+iIVoUcaxWIo7BCdMczwOC+54z0pyC6HVpBCTZR4KpWUTEqPnmqsm6cM9IfPciG7XAR6BIhYGLARRJyyuk3vVIy0j20UJs4jqL5s+1CMXZd4gTQJdsuFNeLA37llVf46KOPaNeuHdHR0cTHx7N27VpiY2NdCmU9evTg97//PeXl5bRt25alS5cSHh5Ofn4+ycnJ5OTkMHz4cNxTUgMDAykuLgbgtdde45NPPqGsrIxp06bx0ksvkZWVxe23387IkSPZtm0bkZGRrF69mi+++IKUlBRmzpyJn58f27dvx8/Pr0bbFy1axLp16/Dz82PZsmV069aNrKwsZs+ezblz52jXrh3vv/8+nTp1qrV95cqVvPTSS1gsFoKDg9mwYQMvvPACJSUlbNmyhfnz51NSUkJKSgpvvPEGs2bNonXr1qSkpHD69GkWLVrE9OnT0XWduXPn8s033xAdHY3NZmP27NlMnz79hn9GV6LFZUFU1Tyw+UlyStrj1G3GqFdaSOy4i1GR29ClYFv+YPbk9EeiVVt04Rr9+mrYewejdwoEIdAKHAjpGYlwhvlWlg+SIP2sCLtuOF8AKSm/qKGXldYaZqhJHEfhHYwICcSmCSyATROMCAm84T53797NqlWrSE9PZ926dR65qeXl5aSkpPD000+7dID37NnDvffey6JFiwB46aWXGDlyJPv372fatGmcOHGi2jXWr1/P0aNH2bVrF2lpaaSmprJp0yYAjh49ypw5c9i/fz8hISGsWrWK6dOnM3jwYJYuXUpaWlqtzhcgODiYH374gblz5/LEE08AhmLbAw88wN69e5k5cya/+tWvrtj+8ssv8+WXX5Kens6aNWvw8fHh5ZdfZsaMGaSlpTFjxoxq183NzWXLli2sXbvWFZb49NNPycrK4sCBAyxZsoTt27dfz4+kzrQ4B5y9J89DTOZyfivCy8EmBZoEC4KevqcAgY673kMFxmjV2DJwxAShRwcYWQtSQhsLQuiV6ma9g3F2DfIQ1pEhPogQHzRhpJtZkHQus9ItMaRW22sSx1F4B4ODA/h7bDd+F9Oh3sIPW7duZcqUKbRq1YqgoCDuvPNO1zF3x3Py5EkmTJhA//79ee2119i/fz8AmzZt4v777wdg0qRJhIZW/8O/fv161q9fT1xcHIMGDeLQoUMcPXoUgK5du7r0fePj48nKyqqT/cnJya73Coe3fft27rvvPgB+9rOfsWXLliu2JyYmMmvWLN5++22cTmfVS9TI1KlT0TSNPn36uKqCbNmyhXvuuQdN04iIiCApKalO93K9tLgQRNUULIDWzlBmFReSZ9Npz0n8L7fmQGlP13Sbe5xXInAG27BcsFeK6zgMVxxEIT3kISa1XoMWOJG1WgwXI1rT49JFys8cJ6hNDEd0Bz00K/YTP9Iz8BIyQGd/oZVOl/Io03PxH7+gVtuvtyabonkwODig0SbfAgIqr1NXHWB3pJTMnz+fRx/11PzIysrC19fXtW+xWCgpKamTje76wtejNQzw5ptvsnPnTr744gvi4+NJTU296mfc7W7qlcAtbgRcsRos1Gcz/YZlEJMYgqY78ROCSK2YSN8Cykr6cLa8sq6cRykgMEIHonI0q7fxAeAirflBxNHqeBJ3FA3grxcC+fCgg8Qd/+DOSz8y5dJWXsjZRcec7/jqTBC5P2Tgm2Gl74nVBOVvJbQgk23brry89XpqsiluThITE/n8888pLS2luLiYtWvX1nhebTrAFSWDANatW0dBQUG1z06YMIH33nvPFQ/OycnhzJkz1c5zJygoiIsXL17V/oq6cStWrGD48OEAjBgxguXLlwOwdOlSRo0adcX2jIwMhg4dyssvv0y7du3Izs6+5uu7k5iYyKpVq9B1nby8PDZu3Finz18vLW4EDNVTsDKPPE+erw2Ezo7W7fm880AsKUbhjZrEcyr0fAHsPSvze7WCcpyhNrLtI0k99zm+Fn/OlJ4gv+wUkZED+MnDc1j87THeWX8YXULXYh27ngsYKWgaEFl6qiFvXXETMWTIECZPnsyAAQMIDw+nf//+BAcHVzuvQgc4NDSUcePG8eOPPwLw4osvkpycTN++fRkxYkSN1cRvu+02Dh486HKQgYGBfPTRR1gslmrnVjBr1ix++ctfXnUSrqCggAEDBuDr68vHH38MwF/+8hd+8Ytf8Nprr7km267U/swzz3D06FGklNx6660MHDiQTp06sXDhQmJjY5k//9oqmN199918/fXX9OnTh+joaAYNGlTjd1nfKDEe4N+fWcQhn2AiLEWc7RHM9oLuWHIue+j3Qm1iOkZ5IJ/d54zCGBr8m/UsEYf/7nGNAeNvp++YcezYuovX9+vk2Npz16WtRF9og71sE6AjNI17X3pVjW69gOYixlNcXExgYCCXL19m9OjRvPXWWwwa1HLL3d8IFd9lfn6+q6J0REREtfPqU4ynRY6A3Uk9XsAKS2+cTgHOSDqIywihu47XJJ7uEa0qd6Lll1WWDtIlOcWFuP/YhKYR3jXGlUI2zWLFOukx+kZOJf1/fg/oCCEY/+Bjyvkq6sQjjzzCgQMHKC0t5YEHHlDO9wb46U9/yoULFygvL+f555+v0fnWNy3aAW9+dzF/2e+H06c9FQsssnPbGFqPeDrbmoTVAaznSnGE+RiZDTpYhU58pC+WPBu604HQNG6d/UuXQLrUdcBBQqsCyC9w04oQlNQxbqVQVMRwmyvTpk1zhTwqePXVV5kwYUITWVQ7jRX3dadFOuDU4wUs/+QLfDNjKPMFV9lhwHKuzOPcGgXULcbyYgEIKRlXvh493gftvJ27eo9g6shfcurIGI9shVNHDroE0t1TyGpqUyhuFj777LOmNqFZ0+IccOrxAma+s4Oy8hAsrXWSLlvRcOKsYYEFVA8/6P4WwqN0zh/VkdLI3+1xtIAhA85SlN2WtlomjDSyFdzDCbWlkKm0MoWi5dLiHPCOzHxD60EInBJKhWRaxFY+yxuOLi0eQuruVAqqB9LJ/yA/dRaTUdabzuWSnoFtyPzuTnTNwrlcJ7C4RrGbqk65tjaFQtEyaHF5wMNi2mLTQEPHpjnoF7yDHJ9AdFnLVyHA0TkAvbUNZztfZJBGqzbnCf/JBibZPmHC4EzsJcJjdZ0SOFcoFNdCi3PAnUr38XTcfzO12xfM6LWS/Z0cbBPDzPpulbhWwEkQl+1oF+1Yzpbhk1LAd4Wj+U/LC1yY0J9RD86ptaKwQtFQJCUl8eWXX3q0vf766zz22GM1nv+HP/zBY3/EiBENZpvi2mlxDvjHjI3cEppBzzZHWX74bjZlDEWcKAU8Mx7cQxHWs6UIaYrm6KBlXMJ5QWdnQRRQfXWd0tpV1ETq8QIWf3uM1OPVV5zVleTkZNfKsAqWL1/u0leoSlUHfCNyk4r6o8XFgI87+pLv04vV5yfg0K1UKuRUUvPSlMosYC2/DK2gjKjYaNdR5XQVV6Ji8rfcoXtUrr5epk+fznPPPUd5eTk+Pj5kZWVx6tQpcnJy6N+/P1JKJk2axKuvvsq8efMoKSkhNjaWvn37snTpUpfcZIU2RFhYGPv27SM+Pp6PPvoIIQT//Oc/eeqppwgICCAxMZHMzMxalzsrro8WNwIu6zmC34sFpLcZgq5V1/d1F98RHu067gWJNB0i2tW/mr/i5qRi8rdq5errpU2bNiQkJLBu3TrAGP2OHz+e3/3ud3zzzTekpaWxe/du/vGPf7Bw4UL8/PxIS0tj6dKl1fras2cPr7/+OgcOHCAzM5OtW7dSWlrKo48+yrp160hNTeXs2bM3ZK+iZlqcAz7fSuAQVvRQP8oHh+FsbRTg9FQ881wBF2U/RWJxKeOLLmPVHQipY9Ed2IrUZJvi2hgW0/aaK1dfK+5hiOXLl9O5c2fGjh1Lu3btsFqtzJw506XdeyUSEhKIiopC0zRiY2PJysri0KFDxMTE0LVrV9e1FPVPiwtBhJ04hpB+cKEcy6nLaEWOGpcbUyFGKSVdykIZ7miLo2QHwcUZ5LTqSMfSXEqOxQOJTXQnCm8ivnMoSx8ads2Vq6+FKVOm8OSTT/L9999z+fJlYmNjycjIqHM/VWUlHQ7HDdumuDZa3AiY/elEHM/FJ+UclpOXXc3uI2BX6EGCFUFnZ2vQnVi0DnQoO8Pgwj10KMuj/5D4RjZe4c3Edw5lTlK3eiusGhgYSFJSErNnzyY5OZmEhAS+++47zp07h9Pp5OOPP2bMmDEA2Gw27Hb7Nffds2dPMjMzXSLrFdKRivqlxTng3v370y4nv1I8B8+MB1GxJ6FHWSn/Hp3LxMFGhkN4WCa+/ndhbTWCVv53YTu6u2luQqEwSU5OJj09neTkZDp06MDChQtJSkpi4MCBxMfHM2XKFMAQ7RkwYAAzZ868pn79/Pz461//ysSJE4mPjycoKKhR5BlbGi1SjvLF//mUD4/7mNNp4AyxQYDNKEJ0sgSEAKmTdCGL9//2uOtzy+a+QIF9FAgL6E5CfTZz3xsvN/j9KJofzUWOsiGpkGeUUjJnzhy6d+/Ok08+2dRmNTn1KUfZ4kbAAHknsqEirxcgzJcu/qcYl7EBqzQm2azSSbijlGVPvQYYymn2y74IqasFF4oWwdtvv+1KXSssLKxWlkhx47S4STiAyMICrIE6DkATcOv+r7nVr5joIf1pv34tJ307Ell6mkjbKAoudeT9h/9EKX3RbRaErhPo3EVMYojK/VXc1Dz55JNqxNvAtEgHPDa+HSU7izhuLaOzw5fbh4Yz6kEjlJB3tICoH0vRfEajWTuClJTp3dCthtaDFGDzL1POt4WjaRp2ux2bzdbUpigaEbvdjqbVX+CgRYYgRj04h9uHFjFJO8HtQ4s8nKn1/F6sPoPRLB2MEvNAq5IDSutB4UGvXr1YsmRJnTILFN6N3W5nyZIl9OrVq976bJGTcFcifdUiil7byPHOd1Li1472ebvp8lBbii4EkL0nj+i4cDX6VXDmzBl+85vfcOjQIXRdv/oHFF6Ppmn06tWLP/7xj7Rv397jmKoJV0c2v7u4RocaHvco2dMj8Nu7mmDHCTo9dA8D7/5tE1qqaI60b9+e//3f/21qMxRejtc4YCHEROC/AQvwjpRy4fX2tfndxezbcQu61oPCHZUC6qczC/nsj7vQ9Y5ogb9k2m8SiIhRuY8KhaJh8IoYsBDCAiwGbgf6AMlCiD7X21/2nrwaBdRTPl2N7gTQ0B3GvkKhUDQUXuGAgQTgmJQyU0pZDiwHplxvZ7UJqNtPfOPRbj/xTf1Yr1AoFDXgFZNwQojpwEQp5UPm/s+AoVLKuW7nPAI8Yu72A/Zdqc/2QYFdfG1+wWX2ksIzF4uzANoGWCPD8I9waq2w6KWc4/Lp/EuOnIa4pysQBpxr5GvWF95qu7faDd5ru7faDTXb3llK2a6uHXlNDPhqSCnfAt4CEEKkXM+MZHNA2d74eKvd4L22e6vdUL+2e0sIIgeIdtuPMtsUCoXCa/EWB7wb6C6E6CqE8AHuBdY0sU0KhUJxQ3hFCEJK6RBCzAW+xEhDe09Kuf8KH3mrcSxrEJTtjY+32g3ea7u32g31aLtXTMIpFArFzYi3hCAUCoXipkM5YIVCoWgibjoHLISYKIQ4LIQ4JoSY1wzseU8IcUYIsc+trY0Q4ishxFHzPdRsF0KIP5u27xVCDHL7zAPm+UeFEA80ku3RQohvhRAHhBD7hRC/9gb7hRCthBC7hBDppt0vme1dhRA7TftWmBO6CCF8zf1j5vEubn3NN9sPCyEmNKTdVe7BIoTYI4RY6022CyGyhBA/CCHShBApZluzfl7M64UIIf4uhDgkhDgohBjeKHZLKW+aF8YEXQYQA/gA6UCfJrZpNDAI2OfWtgiYZ27PA141t+8A1mGUqxsG7DTb2wCZ5nuouR3aCLZ3AAaZ20HAEYyl4M3afvP6gea2Ddhp2vMJcK/Z/ibwmLn9/4A3ze17gRXmdh/zGfIFuprPlqWRnpungGXAWnPfK2wHsoCwKm3N+nkxr/kh8JC57QOENIbdDf4gNeYLGA586bY/H5jfDOzqgqcDPgx0MLc7AIfN7b8ByVXPA5KBv7m1e5zXiPexGviJN9kP+APfA0MxVi9Zqz4rGNk1w81tq3meqPr8uJ/XwDZHAV8D44C1pi3eYnsW1R1ws35egGDgR8ykhMa0+2YLQUQC2W77J8225ka4lDLX3D4NVCi812Z/k9+X+a9tHMZostnbb/4LnwacAb7CGAFekFI6arDBZZ95vBBo2xR2m7wO/BaoEBpui/fYLoH1QohUYcgDQPN/XroCZ4H3zbDPO0KIgMaw+2ZzwF6HNP5UNutcQCFEILAKeEJKWeR+rLnaL6V0SiljMUaTCUD9lTFoQIQQPwXOSClTm9qW62SklHIQhnLhHCHEaPeDzfR5sWKECf9HShkHXMIIObhoKLtvNgfsLUuW84QQHQDM9zNme232N9l9CSFsGM53qZTyU7PZa+yXUl4AvsX4tz1ECFGx+MjdBpd95vFgIJ+msTsRmCyEyMJQ/RuHoYPtDbYjpcwx388An2H88Wvuz8tJ4KSUcqe5/3cMh9zgdt9sDthbliyvASpmSB/AiK1WtP/cnGUdBhSa/wJ9CdwmhAg1Z2JvM9saFCGEAN4FDkop/8tb7BdCtBNChJjbfhhx64MYjnh6LXZX3M904BtzxLMGuNfMNOgKdAd2NZTdAFLK+VLKKCllF4zn9xsp5UxvsF0IESCECKrYxvg576OZPy9SytNAthCip9l0K3CgUexu6KB8Y78wZiiPYMT8nm0G9nwM5AJ2jL+0D2LE6L4GjgIbgDbmuQJDeD4D+AEY7NbPbOCY+fpFI9k+EuPfrr1Amvm6o7nbDwwA9ph27wNeMNtjMJzQMWAl4Gu2tzL3j5nHY9z6eta8n8PA7Y387IylMgui2dtu2phuvvZX/P419+fFvF4skGI+M//AyGJocLvVUmSFQqFoIm62EIRCoVB4DcoBKxQKRROhHLBCoVA0EcoBKxQKRROhHLBCoVA0EcoBK5o1QoipQggphLjqSjYhxBNCCP8buNYsIcQbtbSfNRW+DgghHq7l85NFM1DgU3gPygErmjvJwBbz/Wo8gSG+0xCskMbS5rHAH4QQ4e4HhRBWKeUaKeXCBrq+4iZEOWBFs8XUoBiJsXjlXrd2ixDij0KIfaYe6+NCiF8BHYFvhRDfmucVu31muhDiA3P7TmFo5+4RQmyo6kyvhDSW2GYAnYUQHwgh3hRC7AQWuY+ghRDhQojPhKFJnC6EGGG23y8MreI0IcTfhBCWG/yaFF6McsCK5swU4F9SyiNAvhAi3mx/BEPiM1ZKOQBDp+LPwCkgSUqZdJV+twDDpCG8shxDeeyaEELEYKz4OmY2RQEjpJRPVTn1z8B3UsqBGLoC+4UQvYEZQKI5mnYCM6/12oqbD6+oiqxosSRjCNGA4SiTgVRgPIYIuQNASnm+jv1GAStMgRUfDC3YqzFDCDESKAMelVKeN6QyWCmldNZw/jjg56Z9TqBQCPEzIB7YbX7Wj0qBF0ULRDlgRbNECNEGw4n1F0JIjGonUgjxTB26cV9n38pt+y/Af0kp1wghxgILrqGvFVLKuTW0X6qDPQL4UEo5vw6fUdzEqBCEorkyHVgipewspewipYzGGKmOwhBYf7RCntF01gAXMUonVZAnhOgthNCAaW7twVTKBDZUvbGvgcdM+yxCiGCzbboQon2F3UKIzg10fYUXoBywormSjKEn684qs/0d4ASwVwiRDtxnHn8L+FfFJByGqPZaYBuGIl0FC4CVQohUjBI+DcGvgSQhxA8YYZM+UsoDwHMYFSP2Yvwh6dBA11d4AUoNTaFQKJoINQJWa7xHfwAAADhJREFUKBSKJkI5YIVCoWgilANWKBSKJkI5YIVCoWgilANWKBSKJkI5YIVCoWgilANWKBSKJuL/AC46hxGRyR4zAAAAAElFTkSuQmCC\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"model_trainer.plot_prediction(plot_preds)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "z0EZCKYi3vL9"
},
"source": [
"## Deep learning"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "mNxuMQM4tvp-"
},
"outputs": [],
"source": [
"class DeepTrainer:\n",
" def __init__(self, data, Y_name, \n",
" validation_split=.1, test_size=.3, \n",
" epochs=500, batch_size=8, random_state=24,\n",
" kernel_initializer=keras.initializers.HeNormal(),\n",
" bias_initializer=keras.initializers.HeNormal(),\n",
" verbose=1,show_samples=False): \n",
"\n",
" data_copy = data.copy()\n",
" Y = data_copy[Y_name]\n",
" X = data_copy.drop(Y_name, 1)\n",
"\n",
" self.n_features = len(X.columns)\n",
" \n",
" self.X_train, self.X_test, self.y_train, self.y_test = train_test_split(\n",
" X.values, Y.values, \n",
" test_size=test_size, random_state=random_state)\n",
" \n",
" self.models={}\n",
"\n",
" self.histories={}\n",
" self.evaluate_result={}\n",
" self.training_time={}\n",
" self.metrics={\"RMSE\":{}, \"R2_Score\":{}}\n",
" \n",
" self.validation_split = validation_split\n",
" self.epochs = epochs\n",
" self.batch_size = batch_size\n",
"\n",
" self.kernel_initializer=kernel_initializer\n",
" self.bias_initializer=bias_initializer\n",
"\n",
" if show_samples:\n",
" self.sampling()\n",
"\n",
" def rmse(self, yTrue, yPreds):\n",
" return np.sqrt(mean_squared_error(yTrue, yPreds))\n",
"\n",
" def _eval_model(self, fit_res, name):\n",
" model = self.models[name]\n",
"\n",
" history = fit_res.history\n",
" self.histories[name] = history\n",
" \n",
" evaluate_result = model.evaluate(self.X_test, self.y_test)\n",
" self.evaluate_result[name] = model.evaluate(self.X_test, self.y_test)\n",
"\n",
" yPreds = model.predict(self.X_test) \n",
" self.metrics[\"RMSE\"][name] = self.rmse(self.y_test, yPreds)\n",
" self.metrics[\"R2_Score\"][name] = r2_score(self.y_test, yPreds)\n",
"\n",
" def plot_histories(self, p): \n",
" p(self.histories)\n",
"\n",
" def plot_metrics(self, m): \n",
" m(self.metrics, self.training_time)\n",
"\n",
" def fit(self, model, model_name, show_summary):\n",
" if show_summary: \n",
" model.summary()\n",
" return\n",
"\n",
" self.models[model_name] = model\n",
"\n",
" start = time.process_time()\n",
" fit_res = model.fit(epochs=self.epochs, x=self.X_train, y=self.y_train, \n",
" validation_split=self.validation_split, \n",
" shuffle=True, \n",
" batch_size=self.batch_size, verbose=1) \n",
" end = time.process_time()\n",
" self.training_time[model_name]= round(end - start)\n",
"\n",
" self._eval_model(fit_res, model_name) \n",
"\n",
" # show_summary, no training happens\n",
" def fit_DNN(self, show_summary=False):\n",
" inputs = keras.Input(shape=[self.n_features])\n",
" x = keras.layers.Dense(units=self.n_features//2, \n",
" activation=keras.activations.relu,\n",
" kernel_initializer=self.kernel_initializer,\n",
" bias_initializer=self.bias_initializer,\n",
" kernel_regularizer=keras.regularizers.L1L2(l1=1e-1, l2=1e-1))(inputs)\n",
" outputs = keras.layers.Dense(units=1,\n",
" kernel_initializer=self.kernel_initializer,\n",
" bias_initializer=self.bias_initializer,)(x)\n",
"\n",
" model = keras.models.Model(inputs=[inputs], outputs=[outputs])\n",
"\n",
" model.compile(optimizer=\"adam\", loss=keras.losses.Huber())\n",
" self.fit(model, \"DNN\", show_summary)\n",
"\n",
" # show_summary, no training happens\n",
" def fit_CONV(self, show_summary=False):\n",
" inputs = keras.Input(shape=[self.n_features])\n",
" x = keras.layers.Dense(units=self.n_features//2, \n",
" activation=keras.activations.relu,\n",
" kernel_initializer=self.kernel_initializer,\n",
" bias_initializer=self.bias_initializer,\n",
" kernel_regularizer=keras.regularizers.L1L2(l1=1e-1, l2=1e-1))(inputs)\n",
" x = keras.layers.Lambda(lambda x: tf.expand_dims(x, -1))(x)\n",
" x = keras.layers.Conv1D(2, 2, 1,\n",
" activation=keras.activations.relu,\n",
" kernel_initializer=self.kernel_initializer,\n",
" bias_initializer=self.bias_initializer)(x) \n",
" x = keras.layers.Flatten()(x)\n",
" outputs = keras.layers.Dense(units=1,\n",
" kernel_initializer=self.kernel_initializer,\n",
" bias_initializer=self.bias_initializer,)(x)\n",
"\n",
" model = keras.models.Model(inputs=[inputs], outputs=[outputs])\n",
"\n",
" model.compile(optimizer=\"adam\", loss=keras.losses.Huber())\n",
" self.fit(model, \"Conv\", show_summary)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "crjOhFEi7QOD"
},
"outputs": [],
"source": [
"class DeepPlot:\n",
" def __init__(self, figsize=(15, 5)) -> None:\n",
" self.figsize = figsize\n",
"\n",
" def __call__(self, histories):\n",
" if len(histories) == 0: return\n",
"\n",
" fig, axes = plt.subplots(ncols=len(histories), figsize=self.figsize)\n",
" for idx, (name, history) in enumerate(histories.items()):\n",
" ax = axes if len(histories)==1 else axes[idx]\n",
"\n",
" ax.plot(history[\"loss\"], label=\"Training\")\n",
" ax.plot(history[\"val_loss\"], label=\"Validation\")\n",
" ax.set_ylabel(\"Loss(cost)\")\n",
" ax.set_xlabel(\"Epochs\")\n",
" ax.title.set_text(name)\n",
"\n",
" leg = plt.legend(frameon=True)\n",
" leg.get_frame().set_edgecolor('black')\n",
" leg.get_frame().set_linewidth(1.0)\n",
" plt.suptitle(\"Model performance\")\n",
" plt.tight_layout()\n",
" plt.show() \n",
"\n",
"deep_plot = DeepPlot()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "cZD53pBxC52V"
},
"outputs": [],
"source": [
"class DeepMetricsTable:\n",
" def __init__(self, figsize=(15, 5)) -> None:\n",
" self.figsize = figsize\n",
"\n",
" def __call__(self, metrics, training_time):\n",
" s1 = pd.DataFrame([metrics[\"RMSE\"], \n",
" metrics[\"R2_Score\"]], index=[\"RMSE\", \"R2_Score\"]).T\n",
" display(s1)\n",
"\n",
" s2 = pd.Series(training_time, name=\"Use time(in Sec.)\").to_frame()\n",
" display(s2)\n",
" \n",
" \n",
" colors_ = [c for c, _ in mcolors.CSS4_COLORS.items()]\n",
" color_list=list(map(lambda idx: colors_[(idx+10)%len(colors_)], pd.factorize(s1.index.unique())[0]))\n",
" \n",
" fig, axs=plt.subplots(nrows=2, ncols=1, figsize=self.figsize)\n",
"\n",
" s1[\"RMSE\"].sort_values().plot.bar(color=color_list, rot=360, ax=axs[0], width=.25)\n",
" axs[0].set_title(f\"RMSE\")\n",
" \n",
" s1[\"R2_Score\"].sort_values().plot.bar(color=color_list, rot=360, ax=axs[1], width=.25)\n",
" axs[1].set_title(f\"R2_Score\")\n",
"\n",
" plt.tight_layout()\n",
" plt.show()\n",
" \n",
"deep_metrics_table = DeepMetricsTable()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "4KOJECVWxO39",
"outputId": "41926e6a-ae8f-4b57-b148-b2a8c1d9cfb2"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:11: FutureWarning: In a future version of pandas all arguments of DataFrame.drop except for the argument 'labels' will be keyword-only\n",
" # This is added back by InteractiveShellApp.init_path()\n"
]
}
],
"source": [
"deep_trainer = DeepTrainer(data_feature_cut, \"Price_euros\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "W8gOKFmbBX2R",
"outputId": "537dc9a0-03ed-4390-ffb4-fd3bd63c8c57"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/500\n",
"103/103 [==============================] - 1s 3ms/step - loss: 1105.9323 - val_loss: 1056.5275\n",
"Epoch 2/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1104.1383 - val_loss: 1054.9264\n",
"Epoch 3/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1102.7070 - val_loss: 1053.6145\n",
"Epoch 4/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1101.4818 - val_loss: 1052.4550\n",
"Epoch 5/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1100.4005 - val_loss: 1051.4476\n",
"Epoch 6/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1099.4196 - val_loss: 1050.4340\n",
"Epoch 7/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1098.3680 - val_loss: 1049.2754\n",
"Epoch 8/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1097.1019 - val_loss: 1047.8578\n",
"Epoch 9/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1095.5660 - val_loss: 1046.1268\n",
"Epoch 10/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1093.7654 - val_loss: 1044.1533\n",
"Epoch 11/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1091.7207 - val_loss: 1041.9691\n",
"Epoch 12/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1089.4524 - val_loss: 1039.5426\n",
"Epoch 13/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1086.9652 - val_loss: 1036.9728\n",
"Epoch 14/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1084.2900 - val_loss: 1034.1515\n",
"Epoch 15/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1081.4193 - val_loss: 1031.1649\n",
"Epoch 16/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1078.3699 - val_loss: 1027.9897\n",
"Epoch 17/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1075.1292 - val_loss: 1024.6399\n",
"Epoch 18/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1071.7161 - val_loss: 1021.0776\n",
"Epoch 19/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1068.1302 - val_loss: 1017.3959\n",
"Epoch 20/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1064.3950 - val_loss: 1013.5820\n",
"Epoch 21/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1060.5021 - val_loss: 1009.5892\n",
"Epoch 22/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1056.4324 - val_loss: 1005.4559\n",
"Epoch 23/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1052.2111 - val_loss: 1001.1028\n",
"Epoch 24/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1047.8348 - val_loss: 996.6624\n",
"Epoch 25/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1043.3364 - val_loss: 992.0514\n",
"Epoch 26/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1038.6770 - val_loss: 987.3360\n",
"Epoch 27/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1033.8723 - val_loss: 982.4485\n",
"Epoch 28/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1028.9572 - val_loss: 977.3908\n",
"Epoch 29/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 1023.8870 - val_loss: 972.2485\n",
"Epoch 30/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1018.7130 - val_loss: 966.9221\n",
"Epoch 31/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1013.3654 - val_loss: 961.5593\n",
"Epoch 32/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1007.8970 - val_loss: 956.0164\n",
"Epoch 33/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1002.3058 - val_loss: 950.3679\n",
"Epoch 34/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 996.5894 - val_loss: 944.5744\n",
"Epoch 35/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 990.7587 - val_loss: 938.6961\n",
"Epoch 36/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 984.8107 - val_loss: 932.6619\n",
"Epoch 37/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 978.7132 - val_loss: 926.4818\n",
"Epoch 38/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 972.4910 - val_loss: 920.1891\n",
"Epoch 39/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 966.1620 - val_loss: 913.7735\n",
"Epoch 40/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 959.7111 - val_loss: 907.2810\n",
"Epoch 41/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 953.1600 - val_loss: 900.7565\n",
"Epoch 42/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 946.4933 - val_loss: 893.9136\n",
"Epoch 43/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 939.6981 - val_loss: 887.1619\n",
"Epoch 44/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 932.8037 - val_loss: 880.1094\n",
"Epoch 45/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 925.7720 - val_loss: 873.0123\n",
"Epoch 46/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 918.6396 - val_loss: 865.8462\n",
"Epoch 47/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 911.4050 - val_loss: 858.5355\n",
"Epoch 48/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 904.0565 - val_loss: 851.0703\n",
"Epoch 49/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 896.5815 - val_loss: 843.6344\n",
"Epoch 50/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 889.0170 - val_loss: 835.9722\n",
"Epoch 51/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 881.3348 - val_loss: 828.2219\n",
"Epoch 52/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 873.5252 - val_loss: 820.3138\n",
"Epoch 53/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 865.6794 - val_loss: 812.3873\n",
"Epoch 54/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 857.7260 - val_loss: 804.3978\n",
"Epoch 55/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 849.6760 - val_loss: 796.2283\n",
"Epoch 56/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 841.5460 - val_loss: 787.9626\n",
"Epoch 57/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 833.3726 - val_loss: 779.7659\n",
"Epoch 58/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 825.2833 - val_loss: 771.7866\n",
"Epoch 59/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 817.3937 - val_loss: 763.7526\n",
"Epoch 60/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 809.5347 - val_loss: 756.1491\n",
"Epoch 61/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 801.6733 - val_loss: 748.5037\n",
"Epoch 62/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 793.8380 - val_loss: 740.8939\n",
"Epoch 63/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 786.1335 - val_loss: 733.5429\n",
"Epoch 64/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 778.3575 - val_loss: 726.1984\n",
"Epoch 65/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 770.5702 - val_loss: 718.7515\n",
"Epoch 66/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 762.7074 - val_loss: 711.2415\n",
"Epoch 67/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 754.8882 - val_loss: 703.9768\n",
"Epoch 68/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 747.2263 - val_loss: 696.6694\n",
"Epoch 69/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 739.5430 - val_loss: 689.3608\n",
"Epoch 70/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 731.9026 - val_loss: 682.0883\n",
"Epoch 71/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 724.3529 - val_loss: 675.3057\n",
"Epoch 72/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 716.7922 - val_loss: 668.2878\n",
"Epoch 73/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 709.1924 - val_loss: 661.4882\n",
"Epoch 74/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 701.5800 - val_loss: 655.0396\n",
"Epoch 75/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 694.0278 - val_loss: 648.8134\n",
"Epoch 76/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 686.4951 - val_loss: 642.6276\n",
"Epoch 77/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 679.0031 - val_loss: 636.7357\n",
"Epoch 78/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 671.6328 - val_loss: 631.0949\n",
"Epoch 79/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 664.5078 - val_loss: 625.6432\n",
"Epoch 80/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 657.3851 - val_loss: 620.1736\n",
"Epoch 81/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 650.3035 - val_loss: 614.6207\n",
"Epoch 82/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 643.2968 - val_loss: 608.9368\n",
"Epoch 83/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 636.4802 - val_loss: 603.3163\n",
"Epoch 84/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 629.6476 - val_loss: 597.6912\n",
"Epoch 85/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 622.7589 - val_loss: 591.9796\n",
"Epoch 86/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 616.0082 - val_loss: 586.3674\n",
"Epoch 87/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 609.3514 - val_loss: 580.5125\n",
"Epoch 88/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 602.8085 - val_loss: 574.8559\n",
"Epoch 89/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 596.2446 - val_loss: 569.1282\n",
"Epoch 90/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 589.8007 - val_loss: 563.5533\n",
"Epoch 91/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 583.2778 - val_loss: 558.1536\n",
"Epoch 92/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 576.7305 - val_loss: 552.4621\n",
"Epoch 93/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 570.3133 - val_loss: 547.0137\n",
"Epoch 94/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 564.0743 - val_loss: 541.7193\n",
"Epoch 95/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 557.9985 - val_loss: 536.2923\n",
"Epoch 96/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 552.0277 - val_loss: 531.0306\n",
"Epoch 97/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 546.1353 - val_loss: 525.7528\n",
"Epoch 98/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 540.3164 - val_loss: 520.5783\n",
"Epoch 99/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 534.4869 - val_loss: 515.1569\n",
"Epoch 100/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 528.7158 - val_loss: 509.8262\n",
"Epoch 101/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 522.9138 - val_loss: 504.2603\n",
"Epoch 102/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 517.0449 - val_loss: 499.2961\n",
"Epoch 103/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 511.3488 - val_loss: 494.3476\n",
"Epoch 104/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 505.6450 - val_loss: 489.3654\n",
"Epoch 105/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 500.0517 - val_loss: 484.4193\n",
"Epoch 106/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 494.7399 - val_loss: 479.4276\n",
"Epoch 107/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 489.5126 - val_loss: 474.4555\n",
"Epoch 108/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 484.3678 - val_loss: 469.3954\n",
"Epoch 109/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 479.2165 - val_loss: 464.4249\n",
"Epoch 110/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 474.1432 - val_loss: 459.7624\n",
"Epoch 111/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 469.2130 - val_loss: 455.0815\n",
"Epoch 112/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 464.3550 - val_loss: 450.4687\n",
"Epoch 113/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 459.5803 - val_loss: 445.9127\n",
"Epoch 114/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 454.8640 - val_loss: 441.5643\n",
"Epoch 115/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 450.0576 - val_loss: 437.3647\n",
"Epoch 116/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 445.3280 - val_loss: 433.1728\n",
"Epoch 117/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 440.6296 - val_loss: 429.0633\n",
"Epoch 118/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 436.0440 - val_loss: 425.1172\n",
"Epoch 119/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 431.5069 - val_loss: 421.1255\n",
"Epoch 120/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 426.9822 - val_loss: 416.9219\n",
"Epoch 121/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 422.5858 - val_loss: 413.0117\n",
"Epoch 122/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 418.4445 - val_loss: 408.9632\n",
"Epoch 123/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 414.3763 - val_loss: 404.9876\n",
"Epoch 124/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 410.4980 - val_loss: 401.2143\n",
"Epoch 125/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 406.8586 - val_loss: 397.5127\n",
"Epoch 126/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 403.4611 - val_loss: 393.8421\n",
"Epoch 127/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 400.1816 - val_loss: 390.4115\n",
"Epoch 128/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 397.0133 - val_loss: 387.0685\n",
"Epoch 129/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 393.9970 - val_loss: 383.6797\n",
"Epoch 130/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 391.1696 - val_loss: 380.7136\n",
"Epoch 131/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 388.3519 - val_loss: 377.7066\n",
"Epoch 132/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 385.5999 - val_loss: 374.6908\n",
"Epoch 133/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 383.0142 - val_loss: 371.8631\n",
"Epoch 134/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 380.5389 - val_loss: 369.4521\n",
"Epoch 135/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 378.2347 - val_loss: 366.8288\n",
"Epoch 136/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 376.0343 - val_loss: 364.1517\n",
"Epoch 137/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 373.8521 - val_loss: 361.8264\n",
"Epoch 138/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 371.7953 - val_loss: 359.4394\n",
"Epoch 139/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 369.8094 - val_loss: 357.3528\n",
"Epoch 140/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 367.9090 - val_loss: 355.2993\n",
"Epoch 141/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 366.1790 - val_loss: 353.4476\n",
"Epoch 142/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 364.5188 - val_loss: 351.5639\n",
"Epoch 143/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 362.9078 - val_loss: 349.8529\n",
"Epoch 144/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 361.3503 - val_loss: 348.1083\n",
"Epoch 145/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 359.8799 - val_loss: 346.3943\n",
"Epoch 146/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 358.3719 - val_loss: 344.7066\n",
"Epoch 147/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 356.9724 - val_loss: 343.1950\n",
"Epoch 148/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 355.5202 - val_loss: 341.5725\n",
"Epoch 149/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 354.1268 - val_loss: 340.2340\n",
"Epoch 150/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 352.7318 - val_loss: 338.8098\n",
"Epoch 151/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 351.3411 - val_loss: 337.6195\n",
"Epoch 152/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 349.9412 - val_loss: 336.4145\n",
"Epoch 153/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 348.6187 - val_loss: 335.1821\n",
"Epoch 154/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 347.3148 - val_loss: 333.8946\n",
"Epoch 155/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 346.1159 - val_loss: 332.7007\n",
"Epoch 156/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 344.8676 - val_loss: 331.5227\n",
"Epoch 157/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 343.7257 - val_loss: 330.5146\n",
"Epoch 158/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 342.6182 - val_loss: 329.5385\n",
"Epoch 159/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 341.5904 - val_loss: 328.6941\n",
"Epoch 160/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 340.6340 - val_loss: 327.9831\n",
"Epoch 161/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 339.6826 - val_loss: 327.3299\n",
"Epoch 162/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 338.8059 - val_loss: 326.7053\n",
"Epoch 163/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 337.9749 - val_loss: 326.0151\n",
"Epoch 164/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 337.1589 - val_loss: 325.4156\n",
"Epoch 165/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 336.3952 - val_loss: 324.7047\n",
"Epoch 166/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 335.6400 - val_loss: 324.1151\n",
"Epoch 167/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 334.8905 - val_loss: 323.4838\n",
"Epoch 168/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 334.2055 - val_loss: 322.9155\n",
"Epoch 169/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 333.5276 - val_loss: 322.3004\n",
"Epoch 170/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 332.8465 - val_loss: 321.7180\n",
"Epoch 171/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 332.2268 - val_loss: 321.3422\n",
"Epoch 172/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 331.6306 - val_loss: 320.8562\n",
"Epoch 173/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 331.0362 - val_loss: 320.3304\n",
"Epoch 174/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 330.4718 - val_loss: 319.8080\n",
"Epoch 175/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 329.8936 - val_loss: 319.3361\n",
"Epoch 176/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 329.3653 - val_loss: 318.8383\n",
"Epoch 177/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 328.8324 - val_loss: 318.3750\n",
"Epoch 178/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 328.3731 - val_loss: 317.9249\n",
"Epoch 179/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 327.9188 - val_loss: 317.5041\n",
"Epoch 180/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 327.4962 - val_loss: 317.1481\n",
"Epoch 181/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 327.0242 - val_loss: 316.7694\n",
"Epoch 182/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 326.6555 - val_loss: 316.3757\n",
"Epoch 183/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 326.2196 - val_loss: 316.1022\n",
"Epoch 184/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 325.8327 - val_loss: 315.5767\n",
"Epoch 185/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 325.4481 - val_loss: 315.1955\n",
"Epoch 186/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 325.0767 - val_loss: 314.7803\n",
"Epoch 187/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 324.7232 - val_loss: 314.4075\n",
"Epoch 188/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 324.3387 - val_loss: 313.9253\n",
"Epoch 189/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 323.9486 - val_loss: 313.5293\n",
"Epoch 190/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 323.6200 - val_loss: 313.1563\n",
"Epoch 191/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 323.2921 - val_loss: 312.7831\n",
"Epoch 192/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 322.9776 - val_loss: 312.3882\n",
"Epoch 193/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 322.6621 - val_loss: 312.0663\n",
"Epoch 194/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 322.3672 - val_loss: 311.6424\n",
"Epoch 195/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 322.0883 - val_loss: 311.2184\n",
"Epoch 196/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 321.7879 - val_loss: 310.9028\n",
"Epoch 197/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 321.5135 - val_loss: 310.5868\n",
"Epoch 198/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 321.2265 - val_loss: 310.3674\n",
"Epoch 199/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 320.9427 - val_loss: 310.0471\n",
"Epoch 200/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 320.6936 - val_loss: 309.7670\n",
"Epoch 201/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 320.4101 - val_loss: 309.5229\n",
"Epoch 202/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 320.1313 - val_loss: 309.2873\n",
"Epoch 203/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 319.8544 - val_loss: 309.0421\n",
"Epoch 204/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 319.6461 - val_loss: 308.7783\n",
"Epoch 205/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 319.3134 - val_loss: 308.4886\n",
"Epoch 206/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 319.0606 - val_loss: 308.2067\n",
"Epoch 207/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 318.7951 - val_loss: 307.9530\n",
"Epoch 208/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 318.5992 - val_loss: 307.6446\n",
"Epoch 209/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 318.2945 - val_loss: 307.3668\n",
"Epoch 210/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 318.0609 - val_loss: 307.0905\n",
"Epoch 211/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 317.8137 - val_loss: 306.7748\n",
"Epoch 212/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 317.6131 - val_loss: 306.6242\n",
"Epoch 213/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 317.3457 - val_loss: 306.3637\n",
"Epoch 214/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 317.1286 - val_loss: 306.0288\n",
"Epoch 215/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 316.8638 - val_loss: 305.8002\n",
"Epoch 216/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 316.6237 - val_loss: 305.5744\n",
"Epoch 217/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 316.4032 - val_loss: 305.3055\n",
"Epoch 218/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 316.1309 - val_loss: 305.1051\n",
"Epoch 219/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 315.9059 - val_loss: 304.9274\n",
"Epoch 220/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 315.6499 - val_loss: 304.6887\n",
"Epoch 221/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 315.4127 - val_loss: 304.5298\n",
"Epoch 222/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 315.1927 - val_loss: 304.2905\n",
"Epoch 223/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 314.9239 - val_loss: 304.0657\n",
"Epoch 224/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 314.7211 - val_loss: 303.7892\n",
"Epoch 225/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 314.5232 - val_loss: 303.6472\n",
"Epoch 226/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 314.2931 - val_loss: 303.4251\n",
"Epoch 227/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 314.0804 - val_loss: 303.1817\n",
"Epoch 228/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 313.8297 - val_loss: 302.9886\n",
"Epoch 229/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 313.6404 - val_loss: 302.8240\n",
"Epoch 230/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 313.4284 - val_loss: 302.5828\n",
"Epoch 231/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 313.1988 - val_loss: 302.3858\n",
"Epoch 232/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 312.9843 - val_loss: 302.1498\n",
"Epoch 233/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 312.7737 - val_loss: 301.9515\n",
"Epoch 234/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 312.6054 - val_loss: 301.8451\n",
"Epoch 235/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 312.4037 - val_loss: 301.5805\n",
"Epoch 236/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 312.2304 - val_loss: 301.4458\n",
"Epoch 237/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 312.0457 - val_loss: 301.2226\n",
"Epoch 238/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 311.8817 - val_loss: 301.1269\n",
"Epoch 239/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 311.6567 - val_loss: 300.9912\n",
"Epoch 240/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 311.5089 - val_loss: 300.7639\n",
"Epoch 241/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 311.2979 - val_loss: 300.6158\n",
"Epoch 242/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 311.1220 - val_loss: 300.3908\n",
"Epoch 243/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 310.9485 - val_loss: 300.1930\n",
"Epoch 244/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 310.7788 - val_loss: 300.0109\n",
"Epoch 245/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 310.6016 - val_loss: 299.8626\n",
"Epoch 246/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 310.4245 - val_loss: 299.7051\n",
"Epoch 247/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 310.2499 - val_loss: 299.4994\n",
"Epoch 248/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 310.0487 - val_loss: 299.3495\n",
"Epoch 249/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 309.8475 - val_loss: 299.1742\n",
"Epoch 250/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 309.6951 - val_loss: 298.9968\n",
"Epoch 251/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 309.5363 - val_loss: 298.8336\n",
"Epoch 252/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 309.3465 - val_loss: 298.6310\n",
"Epoch 253/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 309.2068 - val_loss: 298.4906\n",
"Epoch 254/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 309.0297 - val_loss: 298.3168\n",
"Epoch 255/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 308.8947 - val_loss: 298.1223\n",
"Epoch 256/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 308.7326 - val_loss: 298.0311\n",
"Epoch 257/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 308.5772 - val_loss: 297.8370\n",
"Epoch 258/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 308.3899 - val_loss: 297.6739\n",
"Epoch 259/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 308.2254 - val_loss: 297.6172\n",
"Epoch 260/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 308.0917 - val_loss: 297.5181\n",
"Epoch 261/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 307.9373 - val_loss: 297.3381\n",
"Epoch 262/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 307.8031 - val_loss: 297.2492\n",
"Epoch 263/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 307.6649 - val_loss: 297.1451\n",
"Epoch 264/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 307.5462 - val_loss: 297.0545\n",
"Epoch 265/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 307.3809 - val_loss: 296.8765\n",
"Epoch 266/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 307.2356 - val_loss: 296.8333\n",
"Epoch 267/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 307.0699 - val_loss: 296.7125\n",
"Epoch 268/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.9557 - val_loss: 296.5560\n",
"Epoch 269/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.8859 - val_loss: 296.4202\n",
"Epoch 270/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.6930 - val_loss: 296.3400\n",
"Epoch 271/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.5581 - val_loss: 296.2343\n",
"Epoch 272/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.4171 - val_loss: 296.1476\n",
"Epoch 273/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.3331 - val_loss: 296.0383\n",
"Epoch 274/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.2081 - val_loss: 295.9642\n",
"Epoch 275/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 306.0740 - val_loss: 295.7951\n",
"Epoch 276/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 305.9359 - val_loss: 295.7094\n",
"Epoch 277/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 305.8036 - val_loss: 295.6037\n",
"Epoch 278/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 305.6982 - val_loss: 295.5178\n",
"Epoch 279/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 305.5690 - val_loss: 295.4239\n",
"Epoch 280/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 305.4524 - val_loss: 295.2885\n",
"Epoch 281/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 305.3713 - val_loss: 295.1848\n",
"Epoch 282/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 305.1983 - val_loss: 295.1281\n",
"Epoch 283/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 305.0609 - val_loss: 295.0300\n",
"Epoch 284/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 304.9939 - val_loss: 294.8353\n",
"Epoch 285/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 304.8636 - val_loss: 294.7887\n",
"Epoch 286/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 304.7518 - val_loss: 294.6765\n",
"Epoch 287/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 304.6342 - val_loss: 294.5042\n",
"Epoch 288/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 304.5215 - val_loss: 294.4302\n",
"Epoch 289/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 304.4039 - val_loss: 294.4374\n",
"Epoch 290/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 304.2806 - val_loss: 294.2827\n",
"Epoch 291/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 304.1891 - val_loss: 294.2476\n",
"Epoch 292/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 304.0707 - val_loss: 294.1657\n",
"Epoch 293/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 303.9901 - val_loss: 294.0687\n",
"Epoch 294/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 303.8738 - val_loss: 294.0062\n",
"Epoch 295/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 303.7629 - val_loss: 293.9285\n",
"Epoch 296/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 303.6758 - val_loss: 293.8737\n",
"Epoch 297/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 303.5310 - val_loss: 293.7694\n",
"Epoch 298/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 303.4524 - val_loss: 293.5968\n",
"Epoch 299/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 303.3033 - val_loss: 293.5674\n",
"Epoch 300/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 303.2036 - val_loss: 293.5779\n",
"Epoch 301/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 303.1025 - val_loss: 293.4963\n",
"Epoch 302/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 302.9818 - val_loss: 293.5122\n",
"Epoch 303/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 302.9102 - val_loss: 293.3880\n",
"Epoch 304/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 302.7712 - val_loss: 293.2994\n",
"Epoch 305/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 302.6774 - val_loss: 293.2243\n",
"Epoch 306/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 302.6394 - val_loss: 293.1686\n",
"Epoch 307/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 302.4713 - val_loss: 293.1593\n",
"Epoch 308/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 302.3610 - val_loss: 293.0995\n",
"Epoch 309/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 302.2679 - val_loss: 293.0325\n",
"Epoch 310/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 302.1934 - val_loss: 292.9998\n",
"Epoch 311/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 302.0767 - val_loss: 292.8527\n",
"Epoch 312/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 301.9747 - val_loss: 292.8705\n",
"Epoch 313/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 301.8419 - val_loss: 292.8354\n",
"Epoch 314/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 301.8213 - val_loss: 292.7126\n",
"Epoch 315/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 301.6827 - val_loss: 292.5739\n",
"Epoch 316/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 301.5840 - val_loss: 292.5128\n",
"Epoch 317/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 301.4548 - val_loss: 292.4810\n",
"Epoch 318/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 301.3741 - val_loss: 292.4169\n",
"Epoch 319/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 301.2637 - val_loss: 292.3741\n",
"Epoch 320/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 301.2006 - val_loss: 292.4117\n",
"Epoch 321/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 301.0627 - val_loss: 292.3085\n",
"Epoch 322/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 300.9447 - val_loss: 292.3466\n",
"Epoch 323/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 300.9017 - val_loss: 292.2274\n",
"Epoch 324/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 300.7948 - val_loss: 292.3090\n",
"Epoch 325/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 300.7012 - val_loss: 292.2178\n",
"Epoch 326/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 300.5995 - val_loss: 292.2061\n",
"Epoch 327/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 300.5169 - val_loss: 292.0139\n",
"Epoch 328/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 300.4353 - val_loss: 292.1100\n",
"Epoch 329/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 300.3564 - val_loss: 292.0568\n",
"Epoch 330/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 300.2230 - val_loss: 291.9894\n",
"Epoch 331/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 300.1461 - val_loss: 291.9806\n",
"Epoch 332/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 300.0329 - val_loss: 291.9394\n",
"Epoch 333/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 299.9405 - val_loss: 292.0156\n",
"Epoch 334/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 299.8962 - val_loss: 291.8495\n",
"Epoch 335/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 299.7473 - val_loss: 291.7935\n",
"Epoch 336/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 299.6369 - val_loss: 291.6686\n",
"Epoch 337/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 299.5952 - val_loss: 291.5758\n",
"Epoch 338/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 299.4799 - val_loss: 291.4879\n",
"Epoch 339/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 299.3972 - val_loss: 291.4550\n",
"Epoch 340/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 299.3371 - val_loss: 291.3977\n",
"Epoch 341/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 299.2166 - val_loss: 291.3012\n",
"Epoch 342/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 299.1431 - val_loss: 291.3855\n",
"Epoch 343/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 299.0519 - val_loss: 291.3197\n",
"Epoch 344/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 298.9453 - val_loss: 291.0868\n",
"Epoch 345/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 298.9037 - val_loss: 291.1536\n",
"Epoch 346/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 298.8161 - val_loss: 290.9817\n",
"Epoch 347/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 298.7195 - val_loss: 291.0318\n",
"Epoch 348/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 298.6625 - val_loss: 290.8231\n",
"Epoch 349/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 298.5429 - val_loss: 290.8965\n",
"Epoch 350/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 298.4688 - val_loss: 290.7198\n",
"Epoch 351/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 298.3901 - val_loss: 290.7643\n",
"Epoch 352/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 298.3275 - val_loss: 290.6650\n",
"Epoch 353/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 298.2463 - val_loss: 290.5651\n",
"Epoch 354/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 298.1685 - val_loss: 290.5616\n",
"Epoch 355/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 298.1062 - val_loss: 290.4759\n",
"Epoch 356/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 297.9690 - val_loss: 290.4123\n",
"Epoch 357/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 297.8658 - val_loss: 290.3210\n",
"Epoch 358/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.8288 - val_loss: 290.0885\n",
"Epoch 359/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.7198 - val_loss: 290.1019\n",
"Epoch 360/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.6536 - val_loss: 289.9332\n",
"Epoch 361/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.5683 - val_loss: 289.9018\n",
"Epoch 362/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 297.4964 - val_loss: 289.8181\n",
"Epoch 363/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.3984 - val_loss: 289.7887\n",
"Epoch 364/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 297.3152 - val_loss: 289.6747\n",
"Epoch 365/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.2576 - val_loss: 289.6420\n",
"Epoch 366/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 297.1760 - val_loss: 289.6053\n",
"Epoch 367/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.1077 - val_loss: 289.6468\n",
"Epoch 368/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 297.0323 - val_loss: 289.4784\n",
"Epoch 369/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 296.9584 - val_loss: 289.2961\n",
"Epoch 370/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 296.8751 - val_loss: 289.3972\n",
"Epoch 371/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 296.8066 - val_loss: 289.3295\n",
"Epoch 372/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 296.7335 - val_loss: 289.1415\n",
"Epoch 373/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 296.6665 - val_loss: 289.2026\n",
"Epoch 374/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 296.5577 - val_loss: 289.1082\n",
"Epoch 375/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 296.5486 - val_loss: 289.0279\n",
"Epoch 376/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 296.4146 - val_loss: 288.9936\n",
"Epoch 377/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 296.3400 - val_loss: 289.0015\n",
"Epoch 378/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 296.2769 - val_loss: 288.9032\n",
"Epoch 379/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 296.2126 - val_loss: 288.7734\n",
"Epoch 380/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 296.1502 - val_loss: 288.8564\n",
"Epoch 381/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 296.1002 - val_loss: 288.7856\n",
"Epoch 382/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 295.9987 - val_loss: 288.7044\n",
"Epoch 383/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.9277 - val_loss: 288.5870\n",
"Epoch 384/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.8507 - val_loss: 288.5323\n",
"Epoch 385/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 295.7608 - val_loss: 288.4475\n",
"Epoch 386/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 295.6822 - val_loss: 288.4823\n",
"Epoch 387/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.6201 - val_loss: 288.4641\n",
"Epoch 388/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.5851 - val_loss: 288.4196\n",
"Epoch 389/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.4894 - val_loss: 288.3782\n",
"Epoch 390/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 295.4356 - val_loss: 288.3764\n",
"Epoch 391/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.3986 - val_loss: 288.3073\n",
"Epoch 392/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.3201 - val_loss: 288.2633\n",
"Epoch 393/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.2978 - val_loss: 288.0872\n",
"Epoch 394/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.1702 - val_loss: 288.1725\n",
"Epoch 395/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 295.1158 - val_loss: 288.0768\n",
"Epoch 396/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 295.0435 - val_loss: 288.0648\n",
"Epoch 397/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 294.9678 - val_loss: 287.9749\n",
"Epoch 398/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.9150 - val_loss: 288.0111\n",
"Epoch 399/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 294.8323 - val_loss: 287.9470\n",
"Epoch 400/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 294.8101 - val_loss: 287.9087\n",
"Epoch 401/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 294.7094 - val_loss: 287.8864\n",
"Epoch 402/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.6701 - val_loss: 287.7708\n",
"Epoch 403/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.6208 - val_loss: 287.7786\n",
"Epoch 404/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 294.5516 - val_loss: 287.7406\n",
"Epoch 405/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.4803 - val_loss: 287.7148\n",
"Epoch 406/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.3763 - val_loss: 287.5978\n",
"Epoch 407/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.3431 - val_loss: 287.5780\n",
"Epoch 408/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 294.2720 - val_loss: 287.4854\n",
"Epoch 409/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 294.2226 - val_loss: 287.4440\n",
"Epoch 410/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.1029 - val_loss: 287.4290\n",
"Epoch 411/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 294.0691 - val_loss: 287.2512\n",
"Epoch 412/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.9772 - val_loss: 287.2841\n",
"Epoch 413/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.9298 - val_loss: 287.2377\n",
"Epoch 414/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.8536 - val_loss: 287.1697\n",
"Epoch 415/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.7863 - val_loss: 287.1089\n",
"Epoch 416/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.7025 - val_loss: 287.1436\n",
"Epoch 417/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.6724 - val_loss: 286.9636\n",
"Epoch 418/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.5497 - val_loss: 286.9251\n",
"Epoch 419/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.5311 - val_loss: 286.9676\n",
"Epoch 420/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.4419 - val_loss: 286.8822\n",
"Epoch 421/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.3675 - val_loss: 286.7000\n",
"Epoch 422/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.3017 - val_loss: 286.7630\n",
"Epoch 423/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.2260 - val_loss: 286.7450\n",
"Epoch 424/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.1369 - val_loss: 286.6824\n",
"Epoch 425/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.0956 - val_loss: 286.6960\n",
"Epoch 426/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 293.0027 - val_loss: 286.6176\n",
"Epoch 427/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.9613 - val_loss: 286.5386\n",
"Epoch 428/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 292.8896 - val_loss: 286.4779\n",
"Epoch 429/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 292.8367 - val_loss: 286.5313\n",
"Epoch 430/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.7556 - val_loss: 286.3012\n",
"Epoch 431/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.6726 - val_loss: 286.4590\n",
"Epoch 432/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.6271 - val_loss: 286.4126\n",
"Epoch 433/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.5543 - val_loss: 286.3698\n",
"Epoch 434/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.4507 - val_loss: 286.3111\n",
"Epoch 435/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.3949 - val_loss: 286.1843\n",
"Epoch 436/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.3477 - val_loss: 286.3657\n",
"Epoch 437/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.2077 - val_loss: 286.1514\n",
"Epoch 438/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.1187 - val_loss: 286.0854\n",
"Epoch 439/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.0463 - val_loss: 286.0996\n",
"Epoch 440/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 292.0242 - val_loss: 286.0305\n",
"Epoch 441/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 291.9064 - val_loss: 285.9883\n",
"Epoch 442/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 291.8292 - val_loss: 285.9419\n",
"Epoch 443/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 291.7793 - val_loss: 285.7646\n",
"Epoch 444/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 291.7075 - val_loss: 285.7948\n",
"Epoch 445/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 291.6086 - val_loss: 285.7398\n",
"Epoch 446/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.5585 - val_loss: 285.9449\n",
"Epoch 447/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.4609 - val_loss: 285.7340\n",
"Epoch 448/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.3947 - val_loss: 285.7261\n",
"Epoch 449/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.2905 - val_loss: 285.5943\n",
"Epoch 450/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.2419 - val_loss: 285.6697\n",
"Epoch 451/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.1499 - val_loss: 285.5077\n",
"Epoch 452/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.0782 - val_loss: 285.4252\n",
"Epoch 453/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.9714 - val_loss: 285.3764\n",
"Epoch 454/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.9689 - val_loss: 285.2697\n",
"Epoch 455/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 290.8244 - val_loss: 285.1785\n",
"Epoch 456/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 290.7198 - val_loss: 285.1867\n",
"Epoch 457/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.6507 - val_loss: 284.9758\n",
"Epoch 458/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.5785 - val_loss: 285.1049\n",
"Epoch 459/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 290.5274 - val_loss: 285.1425\n",
"Epoch 460/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.4246 - val_loss: 284.9817\n",
"Epoch 461/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.3597 - val_loss: 284.9255\n",
"Epoch 462/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.2699 - val_loss: 284.7197\n",
"Epoch 463/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.2230 - val_loss: 284.8355\n",
"Epoch 464/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.1218 - val_loss: 284.7270\n",
"Epoch 465/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 290.0210 - val_loss: 284.6759\n",
"Epoch 466/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.9581 - val_loss: 284.6264\n",
"Epoch 467/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.8452 - val_loss: 284.6567\n",
"Epoch 468/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.8172 - val_loss: 284.5523\n",
"Epoch 469/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 289.7307 - val_loss: 284.5385\n",
"Epoch 470/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.6617 - val_loss: 284.4145\n",
"Epoch 471/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.5607 - val_loss: 284.4946\n",
"Epoch 472/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.4797 - val_loss: 284.5727\n",
"Epoch 473/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.3810 - val_loss: 284.3777\n",
"Epoch 474/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.3156 - val_loss: 284.2776\n",
"Epoch 475/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.2697 - val_loss: 284.2639\n",
"Epoch 476/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.1695 - val_loss: 284.2200\n",
"Epoch 477/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.1055 - val_loss: 284.1658\n",
"Epoch 478/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.0136 - val_loss: 284.0782\n",
"Epoch 479/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.9139 - val_loss: 284.0251\n",
"Epoch 480/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.9013 - val_loss: 284.0063\n",
"Epoch 481/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.8409 - val_loss: 283.9776\n",
"Epoch 482/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.7112 - val_loss: 283.9574\n",
"Epoch 483/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 288.6124 - val_loss: 284.0346\n",
"Epoch 484/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 288.5884 - val_loss: 284.0312\n",
"Epoch 485/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.5028 - val_loss: 283.9670\n",
"Epoch 486/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.3918 - val_loss: 283.8879\n",
"Epoch 487/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.3226 - val_loss: 283.8732\n",
"Epoch 488/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.2743 - val_loss: 283.7917\n",
"Epoch 489/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.1765 - val_loss: 283.8544\n",
"Epoch 490/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.1697 - val_loss: 283.7698\n",
"Epoch 491/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.0595 - val_loss: 283.7316\n",
"Epoch 492/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 288.0160 - val_loss: 283.6307\n",
"Epoch 493/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.9456 - val_loss: 283.7152\n",
"Epoch 494/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.8442 - val_loss: 283.6199\n",
"Epoch 495/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.8245 - val_loss: 283.6124\n",
"Epoch 496/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.7423 - val_loss: 283.5781\n",
"Epoch 497/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.6516 - val_loss: 283.4628\n",
"Epoch 498/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.6229 - val_loss: 283.4975\n",
"Epoch 499/500\n",
"103/103 [==============================] - 0s 1ms/step - loss: 287.5352 - val_loss: 283.3803\n",
"Epoch 500/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.4958 - val_loss: 283.4253\n",
"13/13 [==============================] - 0s 1ms/step - loss: 300.8203\n",
"13/13 [==============================] - 0s 1ms/step - loss: 300.8203\n"
]
}
],
"source": [
"deep_trainer.fit_DNN(show_summary=False)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "Ou1zzxYo8gvp",
"outputId": "d4db934c-305c-4f1b-ae53-2cbb18776b72"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/500\n",
"103/103 [==============================] - 1s 3ms/step - loss: 1103.1432 - val_loss: 1053.5214\n",
"Epoch 2/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1101.1417 - val_loss: 1051.7905\n",
"Epoch 3/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1099.4484 - val_loss: 1049.9563\n",
"Epoch 4/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1097.2726 - val_loss: 1047.1937\n",
"Epoch 5/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1093.7235 - val_loss: 1042.3810\n",
"Epoch 6/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1087.4648 - val_loss: 1034.0975\n",
"Epoch 7/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1077.3799 - val_loss: 1021.4321\n",
"Epoch 8/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1062.7124 - val_loss: 1003.4720\n",
"Epoch 9/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1042.6041 - val_loss: 979.3077\n",
"Epoch 10/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 1015.5487 - val_loss: 947.0608\n",
"Epoch 11/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 980.8922 - val_loss: 910.4064\n",
"Epoch 12/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 939.8334 - val_loss: 876.0480\n",
"Epoch 13/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 891.5536 - val_loss: 834.9562\n",
"Epoch 14/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 835.2332 - val_loss: 786.7576\n",
"Epoch 15/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 774.4525 - val_loss: 732.0176\n",
"Epoch 16/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 710.9427 - val_loss: 671.4665\n",
"Epoch 17/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 646.3146 - val_loss: 608.4727\n",
"Epoch 18/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 583.2734 - val_loss: 546.5669\n",
"Epoch 19/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 520.5212 - val_loss: 486.9510\n",
"Epoch 20/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 464.6978 - val_loss: 438.7670\n",
"Epoch 21/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 418.4198 - val_loss: 408.5106\n",
"Epoch 22/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 384.2364 - val_loss: 388.5303\n",
"Epoch 23/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 361.0157 - val_loss: 376.9533\n",
"Epoch 24/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 343.4917 - val_loss: 367.0481\n",
"Epoch 25/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 331.9329 - val_loss: 360.1859\n",
"Epoch 26/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 324.3167 - val_loss: 354.5123\n",
"Epoch 27/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 318.4257 - val_loss: 350.9669\n",
"Epoch 28/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 313.6642 - val_loss: 347.3117\n",
"Epoch 29/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 309.9497 - val_loss: 343.5659\n",
"Epoch 30/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 306.5324 - val_loss: 340.3910\n",
"Epoch 31/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 303.7158 - val_loss: 336.6631\n",
"Epoch 32/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 301.1413 - val_loss: 333.6434\n",
"Epoch 33/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 298.4302 - val_loss: 330.7697\n",
"Epoch 34/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 296.1279 - val_loss: 327.5363\n",
"Epoch 35/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 293.9860 - val_loss: 324.6286\n",
"Epoch 36/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 291.6893 - val_loss: 321.7664\n",
"Epoch 37/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 289.5534 - val_loss: 319.5938\n",
"Epoch 38/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 287.5510 - val_loss: 317.6973\n",
"Epoch 39/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 285.7823 - val_loss: 315.6763\n",
"Epoch 40/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 284.0005 - val_loss: 312.9736\n",
"Epoch 41/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 282.1579 - val_loss: 310.7267\n",
"Epoch 42/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 280.4519 - val_loss: 307.9417\n",
"Epoch 43/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 278.9046 - val_loss: 306.3927\n",
"Epoch 44/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 277.4928 - val_loss: 304.6190\n",
"Epoch 45/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 275.9693 - val_loss: 302.2720\n",
"Epoch 46/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 274.5093 - val_loss: 300.7892\n",
"Epoch 47/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 273.0606 - val_loss: 299.2389\n",
"Epoch 48/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 271.5189 - val_loss: 298.2155\n",
"Epoch 49/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 270.0621 - val_loss: 296.7694\n",
"Epoch 50/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 268.7496 - val_loss: 295.4883\n",
"Epoch 51/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 267.6102 - val_loss: 293.9191\n",
"Epoch 52/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 266.4875 - val_loss: 292.3487\n",
"Epoch 53/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 265.5683 - val_loss: 291.4517\n",
"Epoch 54/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 264.6635 - val_loss: 290.6953\n",
"Epoch 55/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 263.7998 - val_loss: 289.7430\n",
"Epoch 56/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 263.0055 - val_loss: 288.6542\n",
"Epoch 57/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 262.2447 - val_loss: 287.6181\n",
"Epoch 58/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 261.3695 - val_loss: 287.5370\n",
"Epoch 59/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 260.6337 - val_loss: 286.3733\n",
"Epoch 60/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 260.0685 - val_loss: 285.5963\n",
"Epoch 61/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 259.4780 - val_loss: 284.8615\n",
"Epoch 62/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 258.9367 - val_loss: 283.5421\n",
"Epoch 63/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 258.1634 - val_loss: 283.1993\n",
"Epoch 64/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 257.7077 - val_loss: 283.2444\n",
"Epoch 65/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 257.2144 - val_loss: 281.8625\n",
"Epoch 66/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 256.7144 - val_loss: 281.7059\n",
"Epoch 67/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 256.1668 - val_loss: 281.5193\n",
"Epoch 68/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 255.6003 - val_loss: 280.3734\n",
"Epoch 69/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 255.0878 - val_loss: 280.3297\n",
"Epoch 70/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 254.6377 - val_loss: 279.8066\n",
"Epoch 71/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 254.0868 - val_loss: 279.5966\n",
"Epoch 72/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 253.6672 - val_loss: 278.4594\n",
"Epoch 73/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 253.2978 - val_loss: 278.5848\n",
"Epoch 74/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 252.7540 - val_loss: 278.4619\n",
"Epoch 75/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 252.2356 - val_loss: 277.6653\n",
"Epoch 76/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 251.9151 - val_loss: 276.7741\n",
"Epoch 77/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 251.5402 - val_loss: 276.6334\n",
"Epoch 78/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 251.2626 - val_loss: 276.7917\n",
"Epoch 79/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 250.8501 - val_loss: 275.9921\n",
"Epoch 80/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 250.6629 - val_loss: 276.1331\n",
"Epoch 81/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 250.4014 - val_loss: 275.0352\n",
"Epoch 82/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 250.1150 - val_loss: 275.0887\n",
"Epoch 83/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 249.8108 - val_loss: 275.3519\n",
"Epoch 84/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 249.6836 - val_loss: 274.5420\n",
"Epoch 85/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 249.5193 - val_loss: 274.2761\n",
"Epoch 86/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 249.2358 - val_loss: 273.8760\n",
"Epoch 87/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 249.1283 - val_loss: 273.6391\n",
"Epoch 88/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 248.8951 - val_loss: 273.7124\n",
"Epoch 89/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 248.6755 - val_loss: 273.6180\n",
"Epoch 90/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 248.3917 - val_loss: 273.3260\n",
"Epoch 91/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 248.3192 - val_loss: 273.1941\n",
"Epoch 92/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 248.1720 - val_loss: 272.4017\n",
"Epoch 93/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 247.8861 - val_loss: 272.6225\n",
"Epoch 94/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 247.6718 - val_loss: 272.5547\n",
"Epoch 95/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 247.5973 - val_loss: 272.2740\n",
"Epoch 96/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 247.4790 - val_loss: 271.2341\n",
"Epoch 97/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 247.1796 - val_loss: 271.5559\n",
"Epoch 98/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 247.0975 - val_loss: 271.5813\n",
"Epoch 99/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 246.9062 - val_loss: 271.2842\n",
"Epoch 100/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 246.7761 - val_loss: 271.5504\n",
"Epoch 101/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 246.5691 - val_loss: 271.0487\n",
"Epoch 102/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 246.4973 - val_loss: 271.1741\n",
"Epoch 103/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 246.3178 - val_loss: 270.4734\n",
"Epoch 104/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 246.2472 - val_loss: 270.3069\n",
"Epoch 105/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 246.1163 - val_loss: 270.3925\n",
"Epoch 106/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.8909 - val_loss: 270.2010\n",
"Epoch 107/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.7391 - val_loss: 270.1645\n",
"Epoch 108/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.6893 - val_loss: 269.7221\n",
"Epoch 109/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.4644 - val_loss: 269.7362\n",
"Epoch 110/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.3856 - val_loss: 269.4934\n",
"Epoch 111/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.1831 - val_loss: 269.3013\n",
"Epoch 112/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.1224 - val_loss: 269.1884\n",
"Epoch 113/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 245.0396 - val_loss: 269.3876\n",
"Epoch 114/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.8040 - val_loss: 269.0448\n",
"Epoch 115/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.7699 - val_loss: 269.1946\n",
"Epoch 116/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.6233 - val_loss: 268.5529\n",
"Epoch 117/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.4972 - val_loss: 268.6808\n",
"Epoch 118/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.3905 - val_loss: 269.3846\n",
"Epoch 119/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.2521 - val_loss: 268.8128\n",
"Epoch 120/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.2237 - val_loss: 268.7630\n",
"Epoch 121/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 244.1534 - val_loss: 268.6448\n",
"Epoch 122/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.8556 - val_loss: 268.1647\n",
"Epoch 123/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.7898 - val_loss: 268.0848\n",
"Epoch 124/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.6810 - val_loss: 268.5864\n",
"Epoch 125/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.5432 - val_loss: 268.1255\n",
"Epoch 126/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.4144 - val_loss: 268.1191\n",
"Epoch 127/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.2628 - val_loss: 268.1647\n",
"Epoch 128/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.1429 - val_loss: 267.9877\n",
"Epoch 129/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 243.2488 - val_loss: 268.2805\n",
"Epoch 130/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.9363 - val_loss: 268.2190\n",
"Epoch 131/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.8306 - val_loss: 268.1108\n",
"Epoch 132/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.7413 - val_loss: 267.8189\n",
"Epoch 133/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.7303 - val_loss: 267.7392\n",
"Epoch 134/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.5983 - val_loss: 267.8636\n",
"Epoch 135/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.4556 - val_loss: 267.9474\n",
"Epoch 136/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.3971 - val_loss: 268.3522\n",
"Epoch 137/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.2655 - val_loss: 267.8976\n",
"Epoch 138/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 242.1996 - val_loss: 267.7425\n",
"Epoch 139/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.9894 - val_loss: 267.7088\n",
"Epoch 140/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.8830 - val_loss: 268.4080\n",
"Epoch 141/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.9166 - val_loss: 267.8352\n",
"Epoch 142/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.7040 - val_loss: 267.8130\n",
"Epoch 143/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.5735 - val_loss: 268.1793\n",
"Epoch 144/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.5536 - val_loss: 268.2153\n",
"Epoch 145/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.4865 - val_loss: 268.1444\n",
"Epoch 146/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.3071 - val_loss: 267.9692\n",
"Epoch 147/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.2545 - val_loss: 267.8131\n",
"Epoch 148/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.2391 - val_loss: 268.1398\n",
"Epoch 149/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.0795 - val_loss: 267.6141\n",
"Epoch 150/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 241.0409 - val_loss: 267.3501\n",
"Epoch 151/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.9026 - val_loss: 267.4209\n",
"Epoch 152/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.8508 - val_loss: 267.5021\n",
"Epoch 153/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.8182 - val_loss: 267.4453\n",
"Epoch 154/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.7439 - val_loss: 267.4299\n",
"Epoch 155/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.7304 - val_loss: 267.2042\n",
"Epoch 156/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.5536 - val_loss: 267.2798\n",
"Epoch 157/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.4631 - val_loss: 267.5405\n",
"Epoch 158/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.4709 - val_loss: 267.4013\n",
"Epoch 159/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.2767 - val_loss: 267.0222\n",
"Epoch 160/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.4097 - val_loss: 266.7014\n",
"Epoch 161/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.1796 - val_loss: 266.9466\n",
"Epoch 162/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.1966 - val_loss: 266.3266\n",
"Epoch 163/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.0393 - val_loss: 266.5672\n",
"Epoch 164/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.0201 - val_loss: 266.2380\n",
"Epoch 165/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 240.1070 - val_loss: 266.2264\n",
"Epoch 166/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.8644 - val_loss: 265.9796\n",
"Epoch 167/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.7285 - val_loss: 266.1887\n",
"Epoch 168/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.7746 - val_loss: 265.6765\n",
"Epoch 169/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.5939 - val_loss: 266.2726\n",
"Epoch 170/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.6214 - val_loss: 266.1961\n",
"Epoch 171/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.5101 - val_loss: 266.0586\n",
"Epoch 172/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.4904 - val_loss: 266.0438\n",
"Epoch 173/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.4168 - val_loss: 266.1636\n",
"Epoch 174/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.3607 - val_loss: 266.2249\n",
"Epoch 175/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.3235 - val_loss: 265.7992\n",
"Epoch 176/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.2154 - val_loss: 265.3040\n",
"Epoch 177/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.2166 - val_loss: 265.3772\n",
"Epoch 178/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.0862 - val_loss: 265.7673\n",
"Epoch 179/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.0677 - val_loss: 265.1934\n",
"Epoch 180/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 239.0543 - val_loss: 265.4024\n",
"Epoch 181/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.9711 - val_loss: 265.4627\n",
"Epoch 182/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.9408 - val_loss: 265.0015\n",
"Epoch 183/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.7591 - val_loss: 265.0363\n",
"Epoch 184/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.6933 - val_loss: 264.8156\n",
"Epoch 185/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.6551 - val_loss: 264.8390\n",
"Epoch 186/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.7557 - val_loss: 264.4837\n",
"Epoch 187/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.7319 - val_loss: 264.3683\n",
"Epoch 188/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.5580 - val_loss: 264.2164\n",
"Epoch 189/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.4902 - val_loss: 264.3723\n",
"Epoch 190/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.3919 - val_loss: 264.1413\n",
"Epoch 191/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.3453 - val_loss: 264.0686\n",
"Epoch 192/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.2106 - val_loss: 264.6616\n",
"Epoch 193/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.2522 - val_loss: 264.1819\n",
"Epoch 194/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.1645 - val_loss: 264.0315\n",
"Epoch 195/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.2447 - val_loss: 263.7386\n",
"Epoch 196/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.2118 - val_loss: 263.3568\n",
"Epoch 197/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.8263 - val_loss: 264.1543\n",
"Epoch 198/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 238.0072 - val_loss: 264.1302\n",
"Epoch 199/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.8327 - val_loss: 263.8028\n",
"Epoch 200/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.9045 - val_loss: 263.7158\n",
"Epoch 201/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.6943 - val_loss: 263.0962\n",
"Epoch 202/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.7955 - val_loss: 262.9698\n",
"Epoch 203/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.6774 - val_loss: 263.0125\n",
"Epoch 204/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.5054 - val_loss: 263.0409\n",
"Epoch 205/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.5199 - val_loss: 263.2270\n",
"Epoch 206/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.4168 - val_loss: 262.6821\n",
"Epoch 207/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.2961 - val_loss: 262.6633\n",
"Epoch 208/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.2607 - val_loss: 263.0221\n",
"Epoch 209/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.2123 - val_loss: 262.6319\n",
"Epoch 210/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.2164 - val_loss: 262.4646\n",
"Epoch 211/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.1128 - val_loss: 262.3473\n",
"Epoch 212/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.0431 - val_loss: 262.3443\n",
"Epoch 213/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 237.0477 - val_loss: 262.4877\n",
"Epoch 214/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.9730 - val_loss: 262.4936\n",
"Epoch 215/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.9304 - val_loss: 262.5242\n",
"Epoch 216/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.9801 - val_loss: 261.7987\n",
"Epoch 217/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.7577 - val_loss: 261.6594\n",
"Epoch 218/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.7008 - val_loss: 261.6725\n",
"Epoch 219/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.6216 - val_loss: 261.6336\n",
"Epoch 220/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.7049 - val_loss: 261.5195\n",
"Epoch 221/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.5621 - val_loss: 261.8341\n",
"Epoch 222/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.4246 - val_loss: 262.2262\n",
"Epoch 223/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.3919 - val_loss: 261.8253\n",
"Epoch 224/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.3862 - val_loss: 261.2649\n",
"Epoch 225/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.2961 - val_loss: 261.6841\n",
"Epoch 226/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.2432 - val_loss: 261.8787\n",
"Epoch 227/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.2295 - val_loss: 261.7922\n",
"Epoch 228/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.0725 - val_loss: 261.3003\n",
"Epoch 229/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 236.1358 - val_loss: 261.1934\n",
"Epoch 230/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.9812 - val_loss: 261.0999\n",
"Epoch 231/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.9093 - val_loss: 261.1083\n",
"Epoch 232/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.8799 - val_loss: 260.8876\n",
"Epoch 233/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.9133 - val_loss: 261.1661\n",
"Epoch 234/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.7521 - val_loss: 260.7428\n",
"Epoch 235/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.7668 - val_loss: 261.2772\n",
"Epoch 236/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.6986 - val_loss: 260.8252\n",
"Epoch 237/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.7516 - val_loss: 260.9157\n",
"Epoch 238/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.5317 - val_loss: 260.4541\n",
"Epoch 239/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.5887 - val_loss: 261.0124\n",
"Epoch 240/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.4220 - val_loss: 260.4262\n",
"Epoch 241/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.4043 - val_loss: 260.2537\n",
"Epoch 242/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.2930 - val_loss: 259.7802\n",
"Epoch 243/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.2492 - val_loss: 259.9637\n",
"Epoch 244/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.2919 - val_loss: 259.9367\n",
"Epoch 245/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.1123 - val_loss: 260.1454\n",
"Epoch 246/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.0714 - val_loss: 259.5034\n",
"Epoch 247/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.0885 - val_loss: 259.3773\n",
"Epoch 248/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.1003 - val_loss: 259.9023\n",
"Epoch 249/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.0686 - val_loss: 259.5245\n",
"Epoch 250/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 235.0604 - val_loss: 259.8058\n",
"Epoch 251/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.8972 - val_loss: 259.3297\n",
"Epoch 252/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.8654 - val_loss: 259.3915\n",
"Epoch 253/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.8125 - val_loss: 259.5095\n",
"Epoch 254/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.6722 - val_loss: 258.6999\n",
"Epoch 255/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.7126 - val_loss: 258.5218\n",
"Epoch 256/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.5013 - val_loss: 258.6403\n",
"Epoch 257/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.5034 - val_loss: 258.6033\n",
"Epoch 258/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.4642 - val_loss: 258.2095\n",
"Epoch 259/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.3062 - val_loss: 258.5686\n",
"Epoch 260/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.4499 - val_loss: 258.4101\n",
"Epoch 261/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.3077 - val_loss: 258.4804\n",
"Epoch 262/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.2738 - val_loss: 257.9120\n",
"Epoch 263/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.1523 - val_loss: 258.1669\n",
"Epoch 264/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.1016 - val_loss: 258.4019\n",
"Epoch 265/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.2078 - val_loss: 257.7454\n",
"Epoch 266/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 234.0568 - val_loss: 258.0580\n",
"Epoch 267/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.9847 - val_loss: 258.1455\n",
"Epoch 268/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.9606 - val_loss: 257.8941\n",
"Epoch 269/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.9084 - val_loss: 257.9042\n",
"Epoch 270/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.7715 - val_loss: 257.1526\n",
"Epoch 271/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.8379 - val_loss: 257.1836\n",
"Epoch 272/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.7848 - val_loss: 257.5616\n",
"Epoch 273/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.5437 - val_loss: 257.2498\n",
"Epoch 274/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.5744 - val_loss: 257.0623\n",
"Epoch 275/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.5076 - val_loss: 257.3199\n",
"Epoch 276/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.4167 - val_loss: 257.1850\n",
"Epoch 277/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.4267 - val_loss: 257.2803\n",
"Epoch 278/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.3318 - val_loss: 257.5933\n",
"Epoch 279/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.3101 - val_loss: 257.0799\n",
"Epoch 280/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.3419 - val_loss: 257.0020\n",
"Epoch 281/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.2644 - val_loss: 256.3699\n",
"Epoch 282/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.1555 - val_loss: 256.6663\n",
"Epoch 283/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.0461 - val_loss: 256.5771\n",
"Epoch 284/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 233.1278 - val_loss: 256.7413\n",
"Epoch 285/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.9798 - val_loss: 256.3734\n",
"Epoch 286/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.9216 - val_loss: 255.8809\n",
"Epoch 287/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.9245 - val_loss: 255.9680\n",
"Epoch 288/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.7835 - val_loss: 256.2813\n",
"Epoch 289/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.8005 - val_loss: 256.1432\n",
"Epoch 290/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.7830 - val_loss: 256.2238\n",
"Epoch 291/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.6280 - val_loss: 256.1407\n",
"Epoch 292/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.4718 - val_loss: 255.8849\n",
"Epoch 293/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.5885 - val_loss: 256.0839\n",
"Epoch 294/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.5192 - val_loss: 255.9634\n",
"Epoch 295/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.2762 - val_loss: 255.5213\n",
"Epoch 296/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.4210 - val_loss: 255.5583\n",
"Epoch 297/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.2617 - val_loss: 255.2538\n",
"Epoch 298/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.2882 - val_loss: 255.3799\n",
"Epoch 299/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.1540 - val_loss: 255.4444\n",
"Epoch 300/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.0935 - val_loss: 255.3695\n",
"Epoch 301/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 232.1126 - val_loss: 255.0173\n",
"Epoch 302/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.9953 - val_loss: 254.6429\n",
"Epoch 303/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.9433 - val_loss: 255.2661\n",
"Epoch 304/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.8948 - val_loss: 254.5799\n",
"Epoch 305/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.6640 - val_loss: 255.1259\n",
"Epoch 306/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.7874 - val_loss: 254.9160\n",
"Epoch 307/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.6641 - val_loss: 254.6109\n",
"Epoch 308/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.5974 - val_loss: 254.4255\n",
"Epoch 309/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.6082 - val_loss: 254.8404\n",
"Epoch 310/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.4984 - val_loss: 254.4444\n",
"Epoch 311/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.3175 - val_loss: 254.4822\n",
"Epoch 312/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.4455 - val_loss: 254.5374\n",
"Epoch 313/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.4619 - val_loss: 254.8899\n",
"Epoch 314/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.2689 - val_loss: 254.3565\n",
"Epoch 315/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.2189 - val_loss: 253.9866\n",
"Epoch 316/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.1839 - val_loss: 253.4600\n",
"Epoch 317/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.1759 - val_loss: 253.7000\n",
"Epoch 318/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.0280 - val_loss: 254.0182\n",
"Epoch 319/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 231.1152 - val_loss: 253.4672\n",
"Epoch 320/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.9487 - val_loss: 253.5712\n",
"Epoch 321/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.9329 - val_loss: 253.8267\n",
"Epoch 322/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.8068 - val_loss: 253.3913\n",
"Epoch 323/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.7624 - val_loss: 253.0479\n",
"Epoch 324/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.6990 - val_loss: 253.3023\n",
"Epoch 325/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.8466 - val_loss: 253.1436\n",
"Epoch 326/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.5635 - val_loss: 253.1807\n",
"Epoch 327/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.6698 - val_loss: 253.4180\n",
"Epoch 328/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.7263 - val_loss: 252.7455\n",
"Epoch 329/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.3765 - val_loss: 252.7106\n",
"Epoch 330/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.5426 - val_loss: 252.8321\n",
"Epoch 331/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.3543 - val_loss: 252.8513\n",
"Epoch 332/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.4336 - val_loss: 252.7334\n",
"Epoch 333/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.2439 - val_loss: 252.4228\n",
"Epoch 334/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.3098 - val_loss: 252.4212\n",
"Epoch 335/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.2680 - val_loss: 252.0584\n",
"Epoch 336/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.1889 - val_loss: 252.1013\n",
"Epoch 337/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 230.2769 - val_loss: 252.3936\n",
"Epoch 338/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.9538 - val_loss: 251.8540\n",
"Epoch 339/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.9246 - val_loss: 251.8034\n",
"Epoch 340/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.8516 - val_loss: 252.0265\n",
"Epoch 341/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.9237 - val_loss: 252.0632\n",
"Epoch 342/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.8058 - val_loss: 251.3709\n",
"Epoch 343/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.8418 - val_loss: 251.1645\n",
"Epoch 344/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.6592 - val_loss: 252.0057\n",
"Epoch 345/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.7520 - val_loss: 251.6664\n",
"Epoch 346/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.7861 - val_loss: 251.8899\n",
"Epoch 347/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.6702 - val_loss: 250.9779\n",
"Epoch 348/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.4819 - val_loss: 251.7213\n",
"Epoch 349/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.5032 - val_loss: 251.5893\n",
"Epoch 350/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.5306 - val_loss: 251.4087\n",
"Epoch 351/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.5137 - val_loss: 251.6195\n",
"Epoch 352/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.3884 - val_loss: 251.2370\n",
"Epoch 353/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.3392 - val_loss: 251.3270\n",
"Epoch 354/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.3972 - val_loss: 251.1413\n",
"Epoch 355/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.2184 - val_loss: 251.0823\n",
"Epoch 356/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.3286 - val_loss: 251.1310\n",
"Epoch 357/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.3400 - val_loss: 250.6827\n",
"Epoch 358/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.1725 - val_loss: 250.6474\n",
"Epoch 359/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.1884 - val_loss: 250.8104\n",
"Epoch 360/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.1937 - val_loss: 251.0257\n",
"Epoch 361/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.1272 - val_loss: 250.1793\n",
"Epoch 362/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.9388 - val_loss: 251.0409\n",
"Epoch 363/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 229.0730 - val_loss: 250.6862\n",
"Epoch 364/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.7790 - val_loss: 250.9642\n",
"Epoch 365/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.9486 - val_loss: 250.6021\n",
"Epoch 366/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.7536 - val_loss: 250.1458\n",
"Epoch 367/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.7828 - val_loss: 250.8556\n",
"Epoch 368/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.8780 - val_loss: 250.5540\n",
"Epoch 369/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.7623 - val_loss: 250.4627\n",
"Epoch 370/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.7062 - val_loss: 250.3239\n",
"Epoch 371/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.6123 - val_loss: 250.5102\n",
"Epoch 372/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.6847 - val_loss: 250.5978\n",
"Epoch 373/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.5497 - val_loss: 250.4841\n",
"Epoch 374/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.5557 - val_loss: 250.4982\n",
"Epoch 375/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.4994 - val_loss: 249.7460\n",
"Epoch 376/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.5080 - val_loss: 250.0738\n",
"Epoch 377/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.5060 - val_loss: 250.2383\n",
"Epoch 378/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.3139 - val_loss: 250.0605\n",
"Epoch 379/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.5397 - val_loss: 249.7799\n",
"Epoch 380/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.3590 - val_loss: 250.3153\n",
"Epoch 381/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.3210 - val_loss: 250.4228\n",
"Epoch 382/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.2747 - val_loss: 249.8180\n",
"Epoch 383/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.1201 - val_loss: 250.2404\n",
"Epoch 384/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.2283 - val_loss: 249.9194\n",
"Epoch 385/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.9533 - val_loss: 250.3818\n",
"Epoch 386/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.0701 - val_loss: 249.9940\n",
"Epoch 387/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 228.0090 - val_loss: 250.3549\n",
"Epoch 388/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.8910 - val_loss: 250.2014\n",
"Epoch 389/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.8546 - val_loss: 250.2784\n",
"Epoch 390/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.8233 - val_loss: 250.3921\n",
"Epoch 391/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.8367 - val_loss: 250.1342\n",
"Epoch 392/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.7159 - val_loss: 250.4074\n",
"Epoch 393/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.6478 - val_loss: 250.5365\n",
"Epoch 394/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.6077 - val_loss: 250.0350\n",
"Epoch 395/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.7339 - val_loss: 250.4097\n",
"Epoch 396/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.5088 - val_loss: 250.1685\n",
"Epoch 397/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.5274 - val_loss: 250.2924\n",
"Epoch 398/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.5838 - val_loss: 250.0820\n",
"Epoch 399/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.3800 - val_loss: 250.0541\n",
"Epoch 400/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.4176 - val_loss: 249.9634\n",
"Epoch 401/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.1892 - val_loss: 250.2318\n",
"Epoch 402/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.2854 - val_loss: 250.1515\n",
"Epoch 403/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.2379 - val_loss: 250.5790\n",
"Epoch 404/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.1009 - val_loss: 250.4124\n",
"Epoch 405/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.1042 - val_loss: 250.3686\n",
"Epoch 406/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.1233 - val_loss: 249.9498\n",
"Epoch 407/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 227.0499 - val_loss: 250.4529\n",
"Epoch 408/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.9516 - val_loss: 250.2952\n",
"Epoch 409/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.9485 - val_loss: 250.2658\n",
"Epoch 410/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.8611 - val_loss: 250.0501\n",
"Epoch 411/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.7657 - val_loss: 250.1589\n",
"Epoch 412/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.8376 - val_loss: 250.1126\n",
"Epoch 413/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.7424 - val_loss: 250.3719\n",
"Epoch 414/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.6208 - val_loss: 250.4906\n",
"Epoch 415/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.7309 - val_loss: 250.4063\n",
"Epoch 416/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.6369 - val_loss: 250.5802\n",
"Epoch 417/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.5517 - val_loss: 250.3090\n",
"Epoch 418/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.5356 - val_loss: 250.3620\n",
"Epoch 419/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.4201 - val_loss: 250.2879\n",
"Epoch 420/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.4890 - val_loss: 249.8669\n",
"Epoch 421/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.2928 - val_loss: 250.2179\n",
"Epoch 422/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.4873 - val_loss: 249.7841\n",
"Epoch 423/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.2980 - val_loss: 250.1381\n",
"Epoch 424/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.1842 - val_loss: 249.9256\n",
"Epoch 425/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.2172 - val_loss: 250.1131\n",
"Epoch 426/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.1908 - val_loss: 250.1975\n",
"Epoch 427/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.0587 - val_loss: 250.1005\n",
"Epoch 428/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.0814 - val_loss: 249.9976\n",
"Epoch 429/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.0802 - val_loss: 250.3387\n",
"Epoch 430/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 226.0181 - val_loss: 250.2142\n",
"Epoch 431/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.9304 - val_loss: 250.3151\n",
"Epoch 432/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.9024 - val_loss: 249.8902\n",
"Epoch 433/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.9386 - val_loss: 249.7493\n",
"Epoch 434/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.9856 - val_loss: 250.1562\n",
"Epoch 435/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.8306 - val_loss: 250.0555\n",
"Epoch 436/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.7941 - val_loss: 249.6305\n",
"Epoch 437/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.7759 - val_loss: 249.8180\n",
"Epoch 438/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.8173 - val_loss: 249.8897\n",
"Epoch 439/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.6756 - val_loss: 250.0273\n",
"Epoch 440/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.6009 - val_loss: 249.7706\n",
"Epoch 441/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.6281 - val_loss: 249.7380\n",
"Epoch 442/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.5820 - val_loss: 249.9315\n",
"Epoch 443/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.7764 - val_loss: 250.5499\n",
"Epoch 444/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.4696 - val_loss: 249.8700\n",
"Epoch 445/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.4611 - val_loss: 249.8495\n",
"Epoch 446/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.4324 - val_loss: 250.2332\n",
"Epoch 447/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.4714 - val_loss: 249.7174\n",
"Epoch 448/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.3874 - val_loss: 249.3525\n",
"Epoch 449/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.5232 - val_loss: 249.6163\n",
"Epoch 450/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.4066 - val_loss: 249.7110\n",
"Epoch 451/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.3924 - val_loss: 250.2359\n",
"Epoch 452/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.2778 - val_loss: 250.0431\n",
"Epoch 453/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.3241 - val_loss: 249.8591\n",
"Epoch 454/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.2062 - val_loss: 250.4048\n",
"Epoch 455/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.1822 - val_loss: 249.9224\n",
"Epoch 456/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.2312 - val_loss: 250.1479\n",
"Epoch 457/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.1185 - val_loss: 249.8516\n",
"Epoch 458/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.1149 - val_loss: 250.0748\n",
"Epoch 459/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.0215 - val_loss: 250.1657\n",
"Epoch 460/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 225.0452 - val_loss: 249.9714\n",
"Epoch 461/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.9596 - val_loss: 249.7849\n",
"Epoch 462/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.9495 - val_loss: 249.8890\n",
"Epoch 463/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.8189 - val_loss: 250.5479\n",
"Epoch 464/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.9650 - val_loss: 250.5702\n",
"Epoch 465/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.7123 - val_loss: 249.8158\n",
"Epoch 466/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.7568 - val_loss: 249.7127\n",
"Epoch 467/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.7155 - val_loss: 250.7606\n",
"Epoch 468/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.6906 - val_loss: 249.9560\n",
"Epoch 469/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.7469 - val_loss: 250.3885\n",
"Epoch 470/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.6768 - val_loss: 250.2869\n",
"Epoch 471/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.7133 - val_loss: 250.2346\n",
"Epoch 472/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.6245 - val_loss: 250.1622\n",
"Epoch 473/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.5910 - val_loss: 250.2397\n",
"Epoch 474/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.6380 - val_loss: 249.7366\n",
"Epoch 475/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.5547 - val_loss: 249.8698\n",
"Epoch 476/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.5692 - val_loss: 249.8485\n",
"Epoch 477/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.3646 - val_loss: 249.7144\n",
"Epoch 478/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.4700 - val_loss: 250.0494\n",
"Epoch 479/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.3142 - val_loss: 250.3940\n",
"Epoch 480/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.2942 - val_loss: 249.7520\n",
"Epoch 481/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.3715 - val_loss: 250.0472\n",
"Epoch 482/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.3476 - val_loss: 250.0172\n",
"Epoch 483/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.2500 - val_loss: 249.9851\n",
"Epoch 484/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.2164 - val_loss: 249.4392\n",
"Epoch 485/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.2696 - val_loss: 250.0448\n",
"Epoch 486/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.3653 - val_loss: 249.6856\n",
"Epoch 487/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.2090 - val_loss: 250.2943\n",
"Epoch 488/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.1509 - val_loss: 249.8417\n",
"Epoch 489/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.1900 - val_loss: 249.7974\n",
"Epoch 490/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.0881 - val_loss: 249.5943\n",
"Epoch 491/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.0699 - val_loss: 249.9142\n",
"Epoch 492/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.0582 - val_loss: 249.6834\n",
"Epoch 493/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.0597 - val_loss: 249.5636\n",
"Epoch 494/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 223.9928 - val_loss: 250.3328\n",
"Epoch 495/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.0345 - val_loss: 249.9331\n",
"Epoch 496/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 224.0071 - val_loss: 249.8002\n",
"Epoch 497/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 223.9769 - val_loss: 250.2679\n",
"Epoch 498/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 223.8831 - val_loss: 249.7563\n",
"Epoch 499/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 223.9494 - val_loss: 249.7729\n",
"Epoch 500/500\n",
"103/103 [==============================] - 0s 2ms/step - loss: 223.7926 - val_loss: 249.9758\n",
"13/13 [==============================] - 0s 1ms/step - loss: 252.4018\n",
"13/13 [==============================] - 0s 1ms/step - loss: 252.4018\n"
]
}
],
"source": [
"deep_trainer.fit_CONV(show_summary=False)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 373
},
"id": "NDWhyqVv6ZXr",
"outputId": "3abad3b4-06c1-4ea5-d079-a009daf81858"
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 1080x360 with 2 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"deep_trainer.plot_histories(deep_plot)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "7Zih2F_sKNKH"
},
"source": [
"## Prediction on unseen test data"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 559
},
"id": "ud_1bPTVFDN-",
"outputId": "f2f5692a-e12e-4623-f82b-896963d8485b"
},
"outputs": [
{
"data": {
"text/html": [
"\n",
" <div id=\"df-4d928dd1-9ec7-4bf1-94e4-e24e528ecf4f\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>RMSE</th>\n",
" <th>R2_Score</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>DNN</th>\n",
" <td>380.818753</td>\n",
" <td>0.729096</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Conv</th>\n",
" <td>359.978705</td>\n",
" <td>0.757935</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4d928dd1-9ec7-4bf1-94e4-e24e528ecf4f')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-4d928dd1-9ec7-4bf1-94e4-e24e528ecf4f button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-4d928dd1-9ec7-4bf1-94e4-e24e528ecf4f');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" RMSE R2_Score\n",
"DNN 380.818753 0.729096\n",
"Conv 359.978705 0.757935"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" <div id=\"df-b13d6704-4aa7-4a36-8102-6196621454bc\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Use time(in Sec.)</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>DNN</th>\n",
" <td>142</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Conv</th>\n",
" <td>172</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b13d6704-4aa7-4a36-8102-6196621454bc')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-b13d6704-4aa7-4a36-8102-6196621454bc button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-b13d6704-4aa7-4a36-8102-6196621454bc');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" Use time(in Sec.)\n",
"DNN 142\n",
"Conv 172"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 1080x360 with 2 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"deep_trainer.plot_metrics(deep_metrics_table)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
},
"id": "yUX7lcJ4v-yQ",
"outputId": "408a7ffa-b094-4946-a1b2-4068e620f483"
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 360x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plot_preds(deep_trainer.models, deep_trainer.X_test, deep_trainer. y_test)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "aXUOIMEQv8yq"
},
"source": [
"## Summary"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "-N_4RDOpt800"
},
"source": [
"- Lasso (L1) can help us figout the useless features and we cut those.\n",
"\n",
"- As for this dataset, ensemble's model is definitely ahead of the general linear model.\n",
"\n",
"- There are many different types of ensemble models, but here most of them maintain an R2-score of 0.8 or higher, which is quite good.\n",
"\n",
"- As you can see from the plot, the KNN is very middle of the road and should make very good predictions.\n",
"\n",
"- The elastic-net, or L1+L2, must be the first choice for the linear model.\n",
"\n",
"- The choice of simple models can be elastic-net, KNN or decision tree for sure.\n",
"\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 746
},
"id": "SH-pPLo1t9eF",
"outputId": "899fd982-98c9-4939-f4aa-34201c1aff66"
},
"outputs": [
{
"data": {
"text/html": [
"\n",
" <div id=\"df-c815955d-8184-41bf-946c-d8d44e23c915\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>RMSE</th>\n",
" <th>R2_Score</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>KNN</th>\n",
" <td>340.583723</td>\n",
" <td>0.783377</td>\n",
" </tr>\n",
" <tr>\n",
" <th>linear_regression</th>\n",
" <td>414.135235</td>\n",
" <td>0.679712</td>\n",
" </tr>\n",
" <tr>\n",
" <th>lasso</th>\n",
" <td>414.155378</td>\n",
" <td>0.679680</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ridge</th>\n",
" <td>414.113387</td>\n",
" <td>0.679745</td>\n",
" </tr>\n",
" <tr>\n",
" <th>elastic_net</th>\n",
" <td>413.999128</td>\n",
" <td>0.679922</td>\n",
" </tr>\n",
" <tr>\n",
" <th>decision_tree</th>\n",
" <td>406.948385</td>\n",
" <td>0.690732</td>\n",
" </tr>\n",
" <tr>\n",
" <th>random_forest</th>\n",
" <td>307.363006</td>\n",
" <td>0.823575</td>\n",
" </tr>\n",
" <tr>\n",
" <th>bagging</th>\n",
" <td>305.001564</td>\n",
" <td>0.826276</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ada_boost</th>\n",
" <td>311.341650</td>\n",
" <td>0.818978</td>\n",
" </tr>\n",
" <tr>\n",
" <th>gradient_boosting</th>\n",
" <td>298.475108</td>\n",
" <td>0.833631</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Voting</th>\n",
" <td>313.899631</td>\n",
" <td>0.815991</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c815955d-8184-41bf-946c-d8d44e23c915')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-c815955d-8184-41bf-946c-d8d44e23c915 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-c815955d-8184-41bf-946c-d8d44e23c915');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" RMSE R2_Score\n",
"KNN 340.583723 0.783377\n",
"linear_regression 414.135235 0.679712\n",
"lasso 414.155378 0.679680\n",
"ridge 414.113387 0.679745\n",
"elastic_net 413.999128 0.679922\n",
"decision_tree 406.948385 0.690732\n",
"random_forest 307.363006 0.823575\n",
"bagging 305.001564 0.826276\n",
"ada_boost 311.341650 0.818978\n",
"gradient_boosting 298.475108 0.833631\n",
"Voting 313.899631 0.815991"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 1080x360 with 2 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"model_trainer.show_metrics(metrics_table)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "llc8xkXxqYGp"
},
"source": [
"- DL works worse than KNN.\n",
"- Convolution 1D (1D conv) works better than the shallow NN.\n",
"- Time consumption of the shallow NN is better than 1D conv "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
},
"id": "XP9mXSloryo5",
"outputId": "e53e2ec5-ccce-4891-b58d-229ee8db8c89"
},
"outputs": [
{
"data": {
"text/html": [
"\n",
" <div id=\"df-397580dc-26e8-4361-911d-61fc5dba6090\">\n",
" <div class=\"colab-df-container\">\n",
" <div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>RMSE</th>\n",
" <th>R2_Score</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>DNN</th>\n",
" <td>380.818753</td>\n",
" <td>0.729096</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Conv</th>\n",
" <td>359.978705</td>\n",
" <td>0.757935</td>\n",
" </tr>\n",
" <tr>\n",
" <th>KNN</th>\n",
" <td>340.583723</td>\n",
" <td>0.783377</td>\n",
" </tr>\n",
" <tr>\n",
" <th>linear_regression</th>\n",
" <td>414.135235</td>\n",
" <td>0.679712</td>\n",
" </tr>\n",
" <tr>\n",
" <th>lasso</th>\n",
" <td>414.155378</td>\n",
" <td>0.679680</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ridge</th>\n",
" <td>414.113387</td>\n",
" <td>0.679745</td>\n",
" </tr>\n",
" <tr>\n",
" <th>elastic_net</th>\n",
" <td>413.999128</td>\n",
" <td>0.679922</td>\n",
" </tr>\n",
" <tr>\n",
" <th>decision_tree</th>\n",
" <td>406.948385</td>\n",
" <td>0.690732</td>\n",
" </tr>\n",
" <tr>\n",
" <th>random_forest</th>\n",
" <td>307.363006</td>\n",
" <td>0.823575</td>\n",
" </tr>\n",
" <tr>\n",
" <th>bagging</th>\n",
" <td>305.001564</td>\n",
" <td>0.826276</td>\n",
" </tr>\n",
" <tr>\n",
" <th>ada_boost</th>\n",
" <td>311.341650</td>\n",
" <td>0.818978</td>\n",
" </tr>\n",
" <tr>\n",
" <th>gradient_boosting</th>\n",
" <td>298.475108</td>\n",
" <td>0.833631</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Voting</th>\n",
" <td>313.899631</td>\n",
" <td>0.815991</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-397580dc-26e8-4361-911d-61fc5dba6090')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-397580dc-26e8-4361-911d-61fc5dba6090 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-397580dc-26e8-4361-911d-61fc5dba6090');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
],
"text/plain": [
" RMSE R2_Score\n",
"DNN 380.818753 0.729096\n",
"Conv 359.978705 0.757935\n",
"KNN 340.583723 0.783377\n",
"linear_regression 414.135235 0.679712\n",
"lasso 414.155378 0.679680\n",
"ridge 414.113387 0.679745\n",
"elastic_net 413.999128 0.679922\n",
"decision_tree 406.948385 0.690732\n",
"random_forest 307.363006 0.823575\n",
"bagging 305.001564 0.826276\n",
"ada_boost 311.341650 0.818978\n",
"gradient_boosting 298.475108 0.833631\n",
"Voting 313.899631 0.815991"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 1440x720 with 2 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"s1 = pd.DataFrame([deep_trainer.metrics[\"RMSE\"], \n",
" deep_trainer.metrics[\"R2_Score\"]], index=[\"RMSE\", \"R2_Score\"]).T\n",
"s2 = pd.DataFrame([model_trainer.metrics[\"RMSE\"], \n",
" model_trainer.metrics[\"R2_Score\"]], index=[\"RMSE\", \"R2_Score\"]).T\n",
" \n",
"summary_df = pd.concat([s1, s2], axis=0)\n",
"display(summary_df)\n",
"\n",
"colors_ = [c for c, _ in mcolors.CSS4_COLORS.items()]\n",
"color_list=list(map(lambda idx: colors_[(idx+10)%len(colors_)], pd.factorize(summary_df.index.unique())[0]))\n",
"\n",
"fig, axs=plt.subplots(nrows=2, ncols=1, figsize=(20, 10))\n",
"\n",
"summary_df[\"RMSE\"].sort_values().plot.bar(color=color_list, rot=360, ax=axs[0], width=.25)\n",
"axs[0].set_title(f\"RMSE\")\n",
"\n",
"summary_df[\"R2_Score\"].sort_values().plot.bar(color=color_list, rot=360, ax=axs[1], width=.25)\n",
"axs[1].set_title(f\"R2_Score\")\n",
"\n",
"plt.tight_layout()\n",
"plt.show()"
]
}
],
"metadata": {
"accelerator": "TPU",
"colab": {
"collapsed_sections": [],
"machine_shape": "hm",
"name": "laptop_price_prediction.ipynb",
"provenance": [],
"toc_visible": true,
"authorship_tag": "ABX9TyOCnxFLj19Q5O+ZXE7g9bhB",
"include_colab_link": true
},
"gpuClass": "standard",
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.13"
},
"vp": {
"vp_config_version": "1.0.0",
"vp_menu_width": 273,
"vp_note_display": false,
"vp_note_width": 0,
"vp_position": {
"width": 278
},
"vp_section_display": false,
"vp_signature": "VisualPython"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment