Last active
March 27, 2020 04:40
-
-
Save kkalucha/f9cf740f5371c15163c2229c701891ce to your computer and use it in GitHub Desktop.
TEST NOTEBOOK
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| attrs==19.3.0 | |
| backcall==0.1.0 | |
| bleach==3.1.4 | |
| cycler==0.10.0 | |
| decorator==4.4.2 | |
| defusedxml==0.6.0 | |
| entrypoints==0.3 | |
| importlib-metadata==1.5.2 | |
| ipykernel==5.2.0 | |
| ipython==7.13.0 | |
| ipython-genutils==0.2.0 | |
| ipywidgets==7.5.1 | |
| jedi==0.16.0 | |
| Jinja2==2.11.1 | |
| joblib==0.14.1 | |
| jsonschema==3.2.0 | |
| jupyter==1.0.0 | |
| jupyter-client==6.1.2 | |
| jupyter-console==6.1.0 | |
| jupyter-core==4.6.3 | |
| kiwisolver==1.1.0 | |
| MarkupSafe==1.1.1 | |
| matplotlib==3.2.1 | |
| mistune==0.8.4 | |
| nbconvert==5.6.1 | |
| nbformat==5.0.4 | |
| notebook==6.0.3 | |
| numpy==1.18.2 | |
| pandas==1.0.3 | |
| pandocfilters==1.4.2 | |
| parso==0.6.2 | |
| patsy==0.5.1 | |
| pexpect==4.8.0 | |
| pickleshare==0.7.5 | |
| prometheus-client==0.7.1 | |
| prompt-toolkit==3.0.5 | |
| ptyprocess==0.6.0 | |
| Pygments==2.6.1 | |
| pyparsing==2.4.6 | |
| pyrsistent==0.16.0 | |
| python-dateutil==2.8.1 | |
| pytz==2019.3 | |
| pyzmq==19.0.0 | |
| qtconsole==4.7.2 | |
| QtPy==1.9.0 | |
| scikit-learn==0.22.2.post1 | |
| scipy==1.4.1 | |
| seaborn==0.10.0 | |
| Send2Trash==1.5.0 | |
| six==1.14.0 | |
| sklearn==0.0 | |
| statsmodels==0.11.1 | |
| terminado==0.8.3 | |
| testpath==0.4.4 | |
| tornado==6.0.4 | |
| traitlets==4.3.3 | |
| wcwidth==0.1.9 | |
| webencodings==0.5.1 | |
| widgetsnbextension==3.5.1 | |
| zipp==3.1.0 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| { | |
| "cells": [ | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "All cases are based on actual LionBase work\n", | |
| "# Case:\n", | |
| "\n", | |
| "FarmX is a company based in Central America that builds bioreactors which can convert wastewater into clean water and algage-based fish feed. Currently, FarmX has dozens of bioreactors. Each day, FarmX collects data on how much feed is produced at each bioreactor, the temperature of the bioreactor, and the nitrate concentration of the water (nitrates stimulate algae growth). They would like to create a model for predicting how much feed is produced at a given bioreactor, as they sell the feed as part of their business model.\n", | |
| "\n", | |
| "To approach this problem, we will use **linear regression** -- a useful and widely used statistical model. **This is often the first type of machine learning that should be tested in any situation.**\n", | |
| "\n", | |
| "Please review Linear Regression - Theory before proceeding with this case.\n", | |
| "\n", | |
| "In this case, we will learn how to:\n", | |
| "- Run a linear regression in Python\n", | |
| "- Interpret linear regression results\n", | |
| "- Identify needs for new features and create them\n", | |
| "- Interpret residual plots\n", | |
| "- Produce an actionable client deliverable\n", | |
| "- Iterate on a model(!!)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 0: Create Dataset for Case" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 1, | |
| "metadata": { | |
| "collapsed": true | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "# Create dataset for problem\n", | |
| "import numpy as np\n", | |
| "import pandas as pd\n", | |
| "import random\n", | |
| "random.seed(30)\n", | |
| "temp_low = np.random.randint(low=-20, high=0, size=300) + np.random.randint(low = 0, high = 100, size = 300)/100\n", | |
| "temp_regular = np.random.randint(low=1, high = 34, size = 800) + np.random.randint(low = 0, high = 100, size = 800)/100\n", | |
| "temp_high = np.random.randint(low= 35, high = 55, size = 300) + np.random.randint(low = 0, high = 100, size = 300)/100\n", | |
| "\n", | |
| "nitrate_low= np.random.randint(low = 20, high = 100, size = 300)/1000 + np.random.randint(low = 0, high = 100, size = 300)/1000\n", | |
| "nitrate_regular = np.random.randint(low = 1, high = 20, size = 800)/1000 + np.random.randint(low = 0, high = 100, size = 800)/1000\n", | |
| "nitrate_high = np.random.randint(low = 20, high = 100, size = 300)/1000 + np.random.randint(low = 0, high = 100, size = 300)/1000\n", | |
| "\n", | |
| "feed_low = np.random.randint(low = 0, high = 5, size = 300) + np.random.randint(low = 0, high = 100, size = 300)/100\n", | |
| "feed_regular = np.random.randint(low=5, high =10, size = 800) + np.random.randint(low = 0, high = 100, size = 800)/100\n", | |
| "feed_high = np.random.randint(low = 0, high =5, size =300) + np.random.randint(low = 0, high = 100, size = 300)/100\n", | |
| "\n", | |
| "all_temp = np.append(temp_low, temp_regular)\n", | |
| "all_temp = np.append(all_temp, temp_high)\n", | |
| "all_nitrate = np.append(nitrate_low, nitrate_regular)\n", | |
| "all_nitrate = np.append(all_nitrate, nitrate_high)\n", | |
| "all_feed = np.append(feed_low,feed_regular)\n", | |
| "all_feed = np.append(all_feed, feed_high)\n", | |
| "\n", | |
| "dataset = pd.DataFrame({'feed': all_feed, 'temperature': all_temp, 'nitrate': all_nitrate})\n", | |
| "\n", | |
| "shuffled_dataset = dataset.sample(frac=1).reset_index(drop=True)\n", | |
| "shuffled_dataset.to_csv('FeedProduction.csv')" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 1: Import Modules" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 2, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "name": "stderr", | |
| "output_type": "stream", | |
| "text": [ | |
| "/Users/kevin_le/anaconda3/lib/python3.6/site-packages/statsmodels/compat/pandas.py:56: FutureWarning: The pandas.core.datetools module is deprecated and will be removed in a future version. Please use the pandas.tseries module instead.\n", | |
| " from pandas.core import datetools\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "import numpy as np\n", | |
| "import pandas as pd\n", | |
| "import scipy as sp\n", | |
| "from sklearn import linear_model\n", | |
| "import statsmodels.api as sm\n", | |
| "import matplotlib.pyplot as plt\n", | |
| "import seaborn as sns" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 2: Read in Data" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 3, | |
| "metadata": { | |
| "collapsed": true | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "df = pd.read_csv('FeedProduction.csv')" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 4, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<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>Unnamed: 0</th>\n", | |
| " <th>feed</th>\n", | |
| " <th>temperature</th>\n", | |
| " <th>nitrate</th>\n", | |
| " </tr>\n", | |
| " </thead>\n", | |
| " <tbody>\n", | |
| " <tr>\n", | |
| " <th>count</th>\n", | |
| " <td>1400.0000</td>\n", | |
| " <td>1400.000000</td>\n", | |
| " <td>1400.000000</td>\n", | |
| " <td>1400.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>mean</th>\n", | |
| " <td>699.5000</td>\n", | |
| " <td>5.274100</td>\n", | |
| " <td>17.424721</td>\n", | |
| " <td>0.080006</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>std</th>\n", | |
| " <td>404.2895</td>\n", | |
| " <td>2.823031</td>\n", | |
| " <td>19.997534</td>\n", | |
| " <td>0.041928</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>min</th>\n", | |
| " <td>0.0000</td>\n", | |
| " <td>0.000000</td>\n", | |
| " <td>-19.850000</td>\n", | |
| " <td>0.003000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>25%</th>\n", | |
| " <td>349.7500</td>\n", | |
| " <td>2.930000</td>\n", | |
| " <td>2.737500</td>\n", | |
| " <td>0.047000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>50%</th>\n", | |
| " <td>699.5000</td>\n", | |
| " <td>5.540000</td>\n", | |
| " <td>16.965000</td>\n", | |
| " <td>0.077000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>75%</th>\n", | |
| " <td>1049.2500</td>\n", | |
| " <td>7.682500</td>\n", | |
| " <td>32.292500</td>\n", | |
| " <td>0.107000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>max</th>\n", | |
| " <td>1399.0000</td>\n", | |
| " <td>9.970000</td>\n", | |
| " <td>54.790000</td>\n", | |
| " <td>0.195000</td>\n", | |
| " </tr>\n", | |
| " </tbody>\n", | |
| "</table>\n", | |
| "</div>" | |
| ], | |
| "text/plain": [ | |
| " Unnamed: 0 feed temperature nitrate\n", | |
| "count 1400.0000 1400.000000 1400.000000 1400.000000\n", | |
| "mean 699.5000 5.274100 17.424721 0.080006\n", | |
| "std 404.2895 2.823031 19.997534 0.041928\n", | |
| "min 0.0000 0.000000 -19.850000 0.003000\n", | |
| "25% 349.7500 2.930000 2.737500 0.047000\n", | |
| "50% 699.5000 5.540000 16.965000 0.077000\n", | |
| "75% 1049.2500 7.682500 32.292500 0.107000\n", | |
| "max 1399.0000 9.970000 54.790000 0.195000" | |
| ] | |
| }, | |
| "execution_count": 4, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df.describe()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 5, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<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>Unnamed: 0</th>\n", | |
| " <th>feed</th>\n", | |
| " <th>temperature</th>\n", | |
| " <th>nitrate</th>\n", | |
| " </tr>\n", | |
| " </thead>\n", | |
| " <tbody>\n", | |
| " <tr>\n", | |
| " <th>0</th>\n", | |
| " <td>0</td>\n", | |
| " <td>5.60</td>\n", | |
| " <td>25.62</td>\n", | |
| " <td>0.102</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>1</th>\n", | |
| " <td>1</td>\n", | |
| " <td>9.87</td>\n", | |
| " <td>29.39</td>\n", | |
| " <td>0.042</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>2</th>\n", | |
| " <td>2</td>\n", | |
| " <td>5.29</td>\n", | |
| " <td>14.92</td>\n", | |
| " <td>0.028</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>3</th>\n", | |
| " <td>3</td>\n", | |
| " <td>9.49</td>\n", | |
| " <td>31.16</td>\n", | |
| " <td>0.031</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>4</th>\n", | |
| " <td>4</td>\n", | |
| " <td>6.01</td>\n", | |
| " <td>31.61</td>\n", | |
| " <td>0.075</td>\n", | |
| " </tr>\n", | |
| " </tbody>\n", | |
| "</table>\n", | |
| "</div>" | |
| ], | |
| "text/plain": [ | |
| " Unnamed: 0 feed temperature nitrate\n", | |
| "0 0 5.60 25.62 0.102\n", | |
| "1 1 9.87 29.39 0.042\n", | |
| "2 2 5.29 14.92 0.028\n", | |
| "3 3 9.49 31.16 0.031\n", | |
| "4 4 6.01 31.61 0.075" | |
| ] | |
| }, | |
| "execution_count": 5, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df.head()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 3: Run a Simple Linear Regression Model\n", | |
| "\n", | |
| "Let's do feed versus nitrate level." | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Using statsmodel" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 6, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<table class=\"simpletable\">\n", | |
| "<caption>OLS Regression Results</caption>\n", | |
| "<tr>\n", | |
| " <th>Dep. Variable:</th> <td>feed</td> <th> R-squared: </th> <td> 0.444</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Model:</th> <td>OLS</td> <th> Adj. R-squared: </th> <td> 0.444</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Method:</th> <td>Least Squares</td> <th> F-statistic: </th> <td> 1118.</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Date:</th> <td>Sat, 21 Mar 2020</td> <th> Prob (F-statistic):</th> <td>1.30e-180</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Time:</th> <td>17:37:39</td> <th> Log-Likelihood: </th> <td> -4079.6</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>No. Observations:</th> <td> 1400</td> <th> AIC: </th> <td> 8161.</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Residuals:</th> <td> 1399</td> <th> BIC: </th> <td> 8167.</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Model:</th> <td> 1</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Covariance Type:</th> <td>nonrobust</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <td></td> <th>coef</th> <th>std err</th> <th>t</th> <th>P>|t|</th> <th>[0.025</th> <th>0.975]</th> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>nitrate</th> <td> 44.1360</td> <td> 1.320</td> <td> 33.433</td> <td> 0.000</td> <td> 41.546</td> <td> 46.726</td>\n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <th>Omnibus:</th> <td>512.682</td> <th> Durbin-Watson: </th> <td> 1.714</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Prob(Omnibus):</th> <td> 0.000</td> <th> Jarque-Bera (JB): </th> <td> 85.970</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Skew:</th> <td>-0.270</td> <th> Prob(JB): </th> <td>2.15e-19</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Kurtosis:</th> <td> 1.912</td> <th> Cond. No. </th> <td> 1.00</td>\n", | |
| "</tr>\n", | |
| "</table>" | |
| ], | |
| "text/plain": [ | |
| "<class 'statsmodels.iolib.summary.Summary'>\n", | |
| "\"\"\"\n", | |
| " OLS Regression Results \n", | |
| "==============================================================================\n", | |
| "Dep. Variable: feed R-squared: 0.444\n", | |
| "Model: OLS Adj. R-squared: 0.444\n", | |
| "Method: Least Squares F-statistic: 1118.\n", | |
| "Date: Sat, 21 Mar 2020 Prob (F-statistic): 1.30e-180\n", | |
| "Time: 17:37:39 Log-Likelihood: -4079.6\n", | |
| "No. Observations: 1400 AIC: 8161.\n", | |
| "Df Residuals: 1399 BIC: 8167.\n", | |
| "Df Model: 1 \n", | |
| "Covariance Type: nonrobust \n", | |
| "==============================================================================\n", | |
| " coef std err t P>|t| [0.025 0.975]\n", | |
| "------------------------------------------------------------------------------\n", | |
| "nitrate 44.1360 1.320 33.433 0.000 41.546 46.726\n", | |
| "==============================================================================\n", | |
| "Omnibus: 512.682 Durbin-Watson: 1.714\n", | |
| "Prob(Omnibus): 0.000 Jarque-Bera (JB): 85.970\n", | |
| "Skew: -0.270 Prob(JB): 2.15e-19\n", | |
| "Kurtosis: 1.912 Cond. No. 1.00\n", | |
| "==============================================================================\n", | |
| "\n", | |
| "Warnings:\n", | |
| "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n", | |
| "\"\"\"" | |
| ] | |
| }, | |
| "execution_count": 6, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "X = df['nitrate']\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "model.summary()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze results\n", | |
| "\n", | |
| "We can interpret the results as follows:\n", | |
| "- The p-value (here described as Prob (F-statistic)) is low meaning that the model is significant\n", | |
| "- The p-value of the Nitrate coefficient (here described as P>|t|) is low meaning that the coefficient is significant\n", | |
| "- The coefficient of Nitrate is positive, meaning that as nitrate levels increase, feed levels increase\n", | |
| "- The Adjusted R^2 value is relatively high, meaning a fair amount of variability is explained by the model\n", | |
| "\n", | |
| "Does this make sense?\n", | |
| "- **This makes sense**, as more nitrates should create more algae feed to be produced.\n", | |
| "\n", | |
| "However, can we keep improving?" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 4: Add more explanatory variables!\n", | |
| "Now this is a multiple linear regression." | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Using statsmodel" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 7, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<table class=\"simpletable\">\n", | |
| "<caption>OLS Regression Results</caption>\n", | |
| "<tr>\n", | |
| " <th>Dep. Variable:</th> <td>feed</td> <th> R-squared: </th> <td> 0.496</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Model:</th> <td>OLS</td> <th> Adj. R-squared: </th> <td> 0.496</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Method:</th> <td>Least Squares</td> <th> F-statistic: </th> <td> 688.8</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Date:</th> <td>Sat, 21 Mar 2020</td> <th> Prob (F-statistic):</th> <td>6.35e-209</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Time:</th> <td>17:37:39</td> <th> Log-Likelihood: </th> <td> -4010.6</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>No. Observations:</th> <td> 1400</td> <th> AIC: </th> <td> 8025.</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Residuals:</th> <td> 1398</td> <th> BIC: </th> <td> 8036.</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Model:</th> <td> 2</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Covariance Type:</th> <td>nonrobust</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <td></td> <th>coef</th> <th>std err</th> <th>t</th> <th>P>|t|</th> <th>[0.025</th> <th>0.975]</th> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>nitrate</th> <td> 32.7974</td> <td> 1.571</td> <td> 20.879</td> <td> 0.000</td> <td> 29.716</td> <td> 35.879</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>temperature</th> <td> 0.0644</td> <td> 0.005</td> <td> 12.037</td> <td> 0.000</td> <td> 0.054</td> <td> 0.075</td>\n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <th>Omnibus:</th> <td>153.946</td> <th> Durbin-Watson: </th> <td> 1.765</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Prob(Omnibus):</th> <td> 0.000</td> <th> Jarque-Bera (JB): </th> <td> 84.477</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Skew:</th> <td>-0.454</td> <th> Prob(JB): </th> <td>4.53e-19</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Kurtosis:</th> <td> 2.210</td> <th> Cond. No. </th> <td> 367.</td>\n", | |
| "</tr>\n", | |
| "</table>" | |
| ], | |
| "text/plain": [ | |
| "<class 'statsmodels.iolib.summary.Summary'>\n", | |
| "\"\"\"\n", | |
| " OLS Regression Results \n", | |
| "==============================================================================\n", | |
| "Dep. Variable: feed R-squared: 0.496\n", | |
| "Model: OLS Adj. R-squared: 0.496\n", | |
| "Method: Least Squares F-statistic: 688.8\n", | |
| "Date: Sat, 21 Mar 2020 Prob (F-statistic): 6.35e-209\n", | |
| "Time: 17:37:39 Log-Likelihood: -4010.6\n", | |
| "No. Observations: 1400 AIC: 8025.\n", | |
| "Df Residuals: 1398 BIC: 8036.\n", | |
| "Df Model: 2 \n", | |
| "Covariance Type: nonrobust \n", | |
| "===============================================================================\n", | |
| " coef std err t P>|t| [0.025 0.975]\n", | |
| "-------------------------------------------------------------------------------\n", | |
| "nitrate 32.7974 1.571 20.879 0.000 29.716 35.879\n", | |
| "temperature 0.0644 0.005 12.037 0.000 0.054 0.075\n", | |
| "==============================================================================\n", | |
| "Omnibus: 153.946 Durbin-Watson: 1.765\n", | |
| "Prob(Omnibus): 0.000 Jarque-Bera (JB): 84.477\n", | |
| "Skew: -0.454 Prob(JB): 4.53e-19\n", | |
| "Kurtosis: 2.210 Cond. No. 367.\n", | |
| "==============================================================================\n", | |
| "\n", | |
| "Warnings:\n", | |
| "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n", | |
| "\"\"\"" | |
| ] | |
| }, | |
| "execution_count": 7, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "X = df[['nitrate', 'temperature']]\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "model.summary()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "We can interpret the results as follows:\n", | |
| "- This model is a slightly better predictor than the previous model as the p-value for the F-statistic is lower than before and the Adj. R-squared value is higher than before. However, the F-statistic itself has decreased.\n", | |
| "- Both Nitrate and Temperature are statistically significant as the p-value is less than 0.05 for both cases\n", | |
| "- Both Nitrate and Temperature are positively correlated with Feed\n", | |
| "\n", | |
| "Does this make sense?\n", | |
| "- This still makes sense. More nitrates and higher temperatures should lead to greater algae feed production. \n", | |
| "\n", | |
| "At this point, we have exhausted all available variables. However, we are still not explaining the variability in the data well. How can we do better?" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 5: Create New Variables" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "Let's see if there's a pattern we're missing in the explanatory variables by plotting temperature and nitrate against each other." | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Using matplotlib.pyplot as plt" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 8, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzsvXuUVNd95/v9VfVpqEYxDTJJpBII\nWXEgZjC0wRITTSZGsY1sWVJfPYxk6S7PHc84uTeZCVjpuchWzCNKIGFsyXdN7kqcxHlJkUFC7kFB\nXpIT4Zl1pSAL1I0IFrKFhICSJiGCxhJdQHX37/5RZ1fvOrVf59Q5VdWwP2shdVefxz6nztm//XsT\nM8Pj8Xg8HhO5dg/A4/F4PJ2PFxYej8fjseKFhcfj8XiseGHh8Xg8HiteWHg8Ho/HihcWHo/H47Hi\nhYXHc5FCRAeJ6GPtHodnauCFhScTiOgIEX283eNIAyLaQEQPt3scacPMi5j5+0A610hEHyOi46kM\nztNxeGHhuaghoq4L4RxZQ1X8fHER4798T+oQ0V8DmAfgSSJ6j4j+S/j5CiJ6nohGiGi/bAIhou8T\n0QPh398joieJ6FIieoSIfkJELxLRfGl7JqL/TESvE9G/ENFWeTIjon9PRK8Q0SkiepqIrozs++tE\n9GMAPw4/+wYRHQvPtY+Ifin8/AYAXwawOhzX/vDzOs1JXpkT0fzwHF8goqMAnrVdf+T+rSOixyOf\nfYOI/p/w538XXve7RPQGEd2tOc4GItpORH8VbnuQiJZLfz9CRB83XOP3ieh3ieg5AKMAPkBE/0d4\nX98Nx/Cr4bYzAHwXwOXhMd4josuJKBdez2Eieiccz2zVeD0dDjP7f/5f6v8AHAHwcen3IoB3AHwa\n1UXKJ8Lf54R//z6A1wBcDWAmgB8C+BGAjwPoAvBXAP5cOh4D2A1gNqqC6UcA/kP4t/7wWL8Q7ns/\ngOcj+34v3LcQfnYPgEvD7e8F8L8ATA//tgHAw5brq20DYH54jr8CMANAwXb9kWNfierk/L7w9zyA\ntwGsCI/3EwALwr9dBmCR5jvYAOBseM48gM0A9qiuQXON3wdwFMCi8L4EAG4MvyMC8MvhOD8Sbv8x\nAMcjx1gDYA+AKwBMA/DHAB5t9/Pp/8X/5zULT6u4B8BTzPwUM08w8/cA7EV1IhP8OTMfZubTqK5S\nDzPz3zHzGIDHAPRFjvn7zHySmY8CeAjAXeHnvwpgMzO/Eu77ewCWytpF+PeTzFwGAGZ+mJnfYeYx\nZv4aqhPbgiaveQMznwnP4XL9CMfyJoCXUBV6AHA9gFFm3hP+PgHgXxFRgZnfZuaDhjH8f+E5xwH8\nNYAlMa/hL5j5YHhfKsy8K/yOmJn/B4BnAPySYf9fBfAVZj7OzOdQFUq3XwimuYsNLyw8reJKAHeE\nJpgRIhoB8G9QXRkL/kn6uaz4/ZLIMY9JP78J4HLpXN+QznMS1ZVwUbMviOje0LxyOtxnJoD3x7rC\nRuRzuFy/zN9gUvh9LvwdzHwGwGoAvwbgbSLaRUQLDWP4X9LPowCmx5yoo/fpU0S0h4hOhtfwaZjv\n05UAviNd8ysAxgH8TIwxeDoALyw8WREtZ3wMwF8zc6/0bwYzb2niHHOln+cBeEs6169GzlVg5udV\n4wv9E/83gM8CmMXMvQBOoypgVNcCAGcA9Ei//6xiG3m/uNf/GICPEdEVAP43hMICAJj5aWb+BKqC\n5hCAP9EcIw668tPyfZoGYAeA/wrgZ8L79BTM9+kYgE9Frns6M5dSGLOnhXhh4cmKfwLwAen3hwHc\nRESriChPRNOpGmp5RRPnGCCiWUQ0F8BvAtgWfv5HAO4jokUAQEQziegOw3F+CsAYgBMAuojoqwDe\nF7mW+VQfDTQM4E4iCkKn8e2Wsca6fmY+garP4M8BvMHMr4TX8jNEdHPoUD4H4D1UV+rNorrGKN2o\nmudOABgjok8B+GTkGJcS0Uzpsz8C8LvCBEhEc4jolhTG62kxXlh4smIzgPtD88NvMfMxALegGnVz\nAtUV5wCaewb/O4B9qE7cuwD8GQAw83cA/D6AbxPRTwD8I4BPGY7zNKo+kh+has46i3rzy2Ph/98h\nopfCn38bVUfvKQAbIa38VSS8/r9B1cEvHzuHqgP+LVTNa78M4P8yndsR1TXWwczvAvjPALajet2f\nA7BT+vshAI8CeD383i8H8I1wm2eI6F1Und3XpjBeT4shZt/8yDP1ICIG8EFmfq3dY/F4Lga8ZuHx\neDweK15YeDwej8eKN0N5PB6Px4rXLDwej8dj5YLJonz/+9/P8+fPb/cwPB6PZ0qxb9++f2HmObbt\nLhhhMX/+fOzdu7fdw/B4PJ4pBRG96bKdN0N5PB6Px4oXFh6Px+Ox4oWFx+PxeKx4YeHxeDweK15Y\neDwej8eKFxYej8fjseKFhcfj8XiseGHh8Xg8HiteWHg8Ho/HihcWHo/H47HihYXH4/F4rHhh4fF4\nPB4rXlh4PB6Px0qmwoKIbiCiV4noNSJap/j7l4joh0T0MhH9PRFdKf3t80T04/Df57Mcp8fj8XjM\nZCYsiCgP4A8BfArAhwDcRUQfimw2BGA5M38YwOMA/iDcdzaA9QCuBXANgPVENCursXo8Ho/HTJaa\nxTUAXmPm15n5PIBvA7hF3oCZdzPzaPjrHgBXhD+vAvA9Zj7JzKcAfA/ADRmO1ePxeDwGshQWRQDH\npN+Ph5/p+AKA78bZl4i+SER7iWjviRMnmhyux+PxeHRkKSxI8RkrNyS6B8ByAFvj7MvM32Tm5cy8\nfM4ca1dAj8fj8SQky7aqxwHMlX6/AsBb0Y2I6OMAvgLgl5n5nLTvxyL7fj+TUXo8Hk9MBodK2Pr0\nq3hrpIzLewsYWLUA/X1F7ecXAsSsXOw3f2CiLgA/AvArAEoAXgTwOWY+KG3Th6pj+wZm/rH0+WwA\n+wB8JPzoJQDLmPmk7nzLly9n34Pb4/FkzeBQCfc9cQDlynjts0KQx23Litixr9Tw+eZbF3e0wCCi\nfcy83LZdZmYoZh4D8BsAngbwCoDtzHyQiDYR0c3hZlsBXALgMSIaJqKd4b4nAfwOqgLmRQCbTILC\n4/F4WsXWp1+tEwgAUK6M49EXjik/3/r0q60cXmZkaYYCMz8F4KnIZ1+Vfv64Yd9vAfhWdqPzeDye\n+Lw1UlZ+Pq6x0ui2n2r4DG6Px+OJweW9BeXneVLF5ei3n2p4YeHxeDwxGFi1AIUgX/dZIcjjrmvn\nKj8fWLWglcPLjEzNUB6Px9PpxI1gEn9T7bP8ytmZRkO1M9oqs2ioVuOjoTweT1x0kU2dGMGU1Vhd\no6G8ZpEiF3KMtcdzIaKLbNr69KuZvrvRuWLlwjnYfeiEce5o11gFXlikRFTql0bKuO+JAwDgBYbH\n06HoIpWyjGBSzRUP7zla+7tu7mjHWGW8gzslTFLf4/F0JrpIpd6eIPaxBodKuG7Ls7hq3S5ct+VZ\nDA6VlNup5oooqrlDN9ZWRVt5YZES7Zb6Hs+FjutkHIeVC9U15U6PVmIdX2gLpZEyGJPageoYrnNC\ndDtdFFaroq28sEiJdkt9j+dCJs5kHIfdh9TVqieAWFaBOJYF1zkhul1/XxGbb12MYm8BBKDYW2ip\nI977LFJiYNUCZaTCwKoF3vHt8WhwfTeycu6aVvlxrAJxLAuquSKKTmPo7yu2be7wmkVK6KQ+gExW\nRB7PVCcN002zZl7TKj+OVSCOZUE1V9yzYl7bNAZXvGaRIiqpf92WZ5Uronu376/t4/FcjMTRFi7v\nLaCkEAzNmnkHVi3AwOP7URmvzzcLchTLF2CyLKhop4aQFK9ZZIyp6NhU0TCycCx6PHFNN1k4d/v7\nith6+xLMkqKfegsBtt6xxDiZR98JAG31J7QCr1lkjG5FBLjZXNvt7/D5I56siKMtmEpsNEvcVb7u\nndh862I8t+567T5T3W/pNYuMUa2IZEw216wiQOLg80c8WRFXW+jvK+K5ddfjjS034rl117dtso37\nTnTCe5wGXrPIGPFA37t9v7LefY4Ig0OllkaAxMHnj3iyIo620Ekrc92zXxop47otz+KtkTJmFgIQ\nASOjFeSIGt591/e4k67bC4sWIL5cVbic8F3I2wk6YaLOyrHo8QBuJqBOM4WaTMvi85FypfZZ0qZI\n9w8ewCN7jkLs3e7r9maoFiBWB7q46rjJO7rP4zqiXbZvd9aox9NpplCbadkV04JrcKhUJygE7bxu\nr1lESFvtU5UVVuGavKObqFWrr7XbhrFm2zBm9QRgBk6XK7VrAuC0WsvSsejxuJCFht3Me656J3Sa\nhg7bgmvr0682CAqBMHe1+j30wkIiidpne+hcioYBzUeAqM4jruPU6KRKLK5pWlfO2R8yFWPCPRcO\naZtC0zBrRd+J+et2WffJE2GC2Uk42QRhO0xSXliE2NQ+ndPN9tC5rH7SSN6Js8oqV8a1Asw7rj2d\nRtyENxtZBI7M6gnqFmVR4jYpctFWWh3s4oVFiEntkydQWZNwiXLQfelxVhlRVNpMElVYhXdcezqN\ntE2hWZi11t+0SJkJDlQT9OKOd2DVAqzdNqydkwStXNx5YRFiuuliAo1qEi5RDrpVUdLsTp02c9uy\nInbsKzmZvIDqSuhsZcLZHxJ9UQHvx/C0jjRNoVlE+KUt0Pr7ilizbdi6XSsXd15YhOgeIALqJkdX\n/4M8wc4sBJge5DAyWmn6IdKp0LsPncDmWxdj69OvojRSBgHaVUkhyGP9TYtqxzM93CrhNPDYfoBQ\nW0W1O6TPM/WJ63BuxkGtWsARmnccp+3bK1qsBa2OSvTCIkT3AN29Yl5s/8PKhXPqjjVSrqAQ5PHg\n6qVNP0ymhKC124ZxeW8BD61eCmBSEPQqoqHEOGzjUQmnykSjGGq1/dRz4RDX4dysg1rWAqILK9dj\n2YTV4FAJG3YerOVbzOoJsP6mRU7jE8fWLV4ZyUxbzeKFRYiLGunqf0jDgaZ7GE2+CbmUgKlOTRzS\nqOnv8ZiI+76k8X4JLeC6Lc82vE+2Y9mE1eBQCQOP7a9bVJ0arWDgcXulaVuofbsEBeCFRR02NdLV\n/7BWY2t0nUxND6NL45Q0V/lxHOfeOe5Jgkv5DHnBlKaDOsmxbMJq69OvKrXvyjjXEup0i1IXU3e7\nzL4+gzsGrm0Nm22xansY5THoSGuV75qt6rO6PUkxvRdy8b2Bx/ZjcKiUWgvjwaEScqR+i3p7Am11\nA5uAMb174jp0RQVd39t2ZHJ7zSImLk6slQvnNORsxIk0sj2M8hhUajSQ3ipfnGfjkwcb4sjbaT/1\nXDi4aMtA1Ve2YedBbLh5Uey8i+h7tnLhHOzYV1JGNOZzhPfOjtWe9+hKfmYhqKv9JJhZCGr/V/1d\nvg4ZeSEYR5MvjZS1RUizwGsWKTM4VMKOfaU6QUEAblvWKGRUpYvXbBuGTmVQCYCktZvi1JHq7yui\np7txXSEERTvLRXumPv19Rdy2rIh8uMo3acwj5Yqzhi9QvWeP7DmqFU7jE6yd0AFAo4yAqHquM+fH\nTJerRAiIuHWnWlnq3GsWKaMru7H70ImGbTc+eVD5wKrSN0wN3MV5XcIIo1EagJsNtBMq4HouTMQC\nS6zybYloQLwwVVMpnDgIH4ouU3tktFL1VygS82wQUKcliGiofJj4qwuFb2UUohcWCpqJ4XadVAeH\nSsbyAIB7lrd4ccS4124bxtanX1WG8+nUfdtD50uVe7LCNX8JQF37U1fSXNCYcpgu7y0kPhcDtfdP\nJQgHh0raJL1WLdi8GSpCs12tXJ1vLs6pCWbnrmAu47a9lGLl5FqqnFD1z3g8zeA62QV5qiWTxkH3\nTprMXSZYsa/Ir9I5zF0w3Yf+viKKKTn2k+KFRQSX2vkme7+rD8HlBYnzELiM2+WcIlqjb9Mziuur\nX08xgB37SlOuPaSns3B9zld/dG4ic4vunbx7xTz0FvSaimnaF/464TMR5XZ0JYDkYxYC9bRruw86\nf8bo+bGWvIPeDBXBZkZSlr94fD827DxYy5D+yLyZ2PP6KYwzI0+kdG7boh5cndTCXOZSBNE10qIy\nwXWRIAOP7wdYn7l97/b9texxHxXliYtrNNSOfSUsv3J27OdL5QcoV8bx6AvHMM6MWT0BRkYrDe8Q\nAzWfQRQR2CG4bsuzTqY0BjA9yGNsvN6JHuSo9r7rzODiOqI+x1OjlZbkXXjNIoLNjKQsfzHOGClX\nauaf5w6frD1g48zK1bct6uEj82Y6ZXqWDIICqD6c89ftwtKNz+DUmXOGLfVUxhujQ2TGmad0I3pP\nexHRTTZ/RDO5Bf19xdo7J7+bQHWy1T3d48ypWQoEp0YrDWpLZYKxZtsw+jY9Y8zD6O8rYsa0xjV+\nK/IuMhUWRHQDEb1KRK8R0TrF3/8tEb1ERGNEdHvkb+NENBz+25nVGKMmpZUL5xgfjiTOJNUXKcIF\ndTx/+KRy0hXjXbNt2NkpCFRDDkcrE3WfzeoJjGp4EtrZ9tEzddGFZ0dpxpkbx5EuEGG5SRNxVeSJ\ntBFTp0YrxrBdoH2RiZkJCyLKA/hDAJ8C8CEAdxHRhyKbHQXw7wD8jeIQZWZeGv67OYsxqpzCO/aV\ncNuySWeSUFm3Pv2qMXvUhioaasc+/QpcREfoxpsGPd1d2HDzolT6Ccv4cFpPEtL24yU5voxwWrtE\nRqosBUGeEOTqVQhZs0k69rQy2OOSpWZxDYDXmPl1Zj4P4NsAbpE3YOYjzPwygAnVAbJG5xR+eM9R\njJ4fQ5CbtFcKdTBp9I8qGipuX+4kKyPb8aMJTr2FAEHeLaIjFyN50OOxYXtuZLt+FscHqs9/1Gnt\nEhmpShTcevsSbL1jSYNWootqch170kTcZsnSwV0EcEz6/TiAa2PsP52I9gIYA7CFmQejGxDRFwF8\nEQDmzZsXe4CmlYYqB0L0jbC1UIySVjRU2it2cfxoXLepRLJAlEFPs92l5+JmYNWChmqtMpdM72rK\ngeviSJ8xrQvD6z8JQO20NuUj6RIFVZ+5OPQF0Xcq7UZLrmQpLFTrzjj61zxmfouIPgDgWSI6wMyH\n6w7G/E0A3wSA5cuXx9btkrQifWukjAdXL1X2vhDhdCsXzsHuQyeMX2SSaCjTPsXeAkbPjzkLMbmp\nUxTx0PdtekZ5vGJvoe56fMc8T2oYlNoRx2dbXuyIaCZRv2zzrYsboolk5AVZmr6BaITTbcuK2H3o\nhHX+0ZUKSrvRkgtZCovjAOZKv18B4C3XnZn5rfD/rxPR9wH0AThs3CkmKxfOwcN7jsba53Jpomxm\nkrT12J3W1WghtJVIHxwqafsAy0SbOqkYHCrhvbONNW5yhIZVjhcOnjSwlcpwMSPpWh/LfV5mTOvS\nCgv5HGlVLVCF2+/YV8LmWxcDAL60bVhrh9eVCmoHWQqLFwF8kIiuAlACcCeAz7nsSESzAIwy8zki\nej+A6wD8QdoDjPslyKt9lelGVXtfR39fEXvfPNlQnVYwUm6MnbYJqf6+onbV5Fo6xGaCmmBg75sn\nvYDwpI5pxe5q3jT59YQJyXQe+Ry6xdnAqgVG7SX6buh8oxufPIie7i6rw7ZTAkYyExbMPEZEvwHg\naQB5AN9i5oNEtAnAXmbeSUQfBfAdALMA3EREG5l5EYBfAPDHRDSBqhN+CzP/MO0x2r6EIE+Y0d2l\nbEcqo1o5rNk2XCunrJtYH+hfjOVXztZOzir7qG4lLx5e3YpJlA4xYevSJXhkz9FEyVEejwlTJ0pT\nVVkZ2zstFlmq8/QWgoZ3DWhcnAEwai/yvqYxnRqtOJmNZ6Yc3p6UTDO4mfkpAE9FPvuq9POLqJqn\novs9D2BxlmMD7D4AlxX4WyNl5DRZnirtIIqY/K9at0upYbisKlwmeRfV2TXaSi565vGkhWsnSpmo\nL8DWS0JM+KrzbLi5se6UanFmytZWLfBM84wuQ1ymiXJTqXJRl/vQRUfM6gmMguL+wQN15iPTl+1a\nQjiJfdQlaglwV+HjOPs7RTX2XDgkKbcf1eiDPCFHVXNpFPEeNOtztD374j1yeT9FhrhpkRYn8jJL\nLmphkaTWyuBQSetn0OEysZrsoypcTUZxuti5rHIEPpfCkwXN9qnQOch7C0GdSdi1rL8KWyRjjoBf\n+O3volyxp4+J99MkVPIdolpc1MICQK3BelR1lVPs5RXI6Pmx2I1TXCbWOKudwaES7t2+3zqxR4ud\nRY8RPZeroPC5FJ5OII52O2NaY47G4FCpLq9DVFwGzAX5bPkaEwwnQRHkCKPnx2pFOHVE38tm+u00\nw0UvLAD9QyccVrKaa0Kl/saZWF1WVUKjsE3spmxXlfp+3xMH0Guw9/p+255Ow+afkFG94xt2HmxI\nAKxMML78xMvWSEYATgs2FYTq2M+cr+/zrWuqJGd8695deVxZ4YUFzGqla5YlAfj6Z6tZzUmkvrxa\nmFkIQFRNQooew9UJbcp21YXy5QhK++msngDrb9JHdXk8aWNbPcftda1auesEzWhloq7FqQrxtziZ\n2MCktn/dlmcbzi+aKskCQ9SnEmH5qmCaVrVW9cIC7vX0dUST3FTqrikmO6oOR/tjy/0yXNcxpmxX\nnSZ15vw47lkxz5p97vFkicvqWZfAN6M7H5qBmitBc+92vTmqYWEHrlV0JgJ0yoZcNUG3OBXau3j/\nVi6cgx37Sg1hulFaEXBy0QsL8cWXK+PODt7eQoAZ07qcfAsbnzxYF82gislWqcMyol9GHC6PqK7i\n4e7tCYwCZ/ehE1o/h8fTCkxdH8V7ppscR8+P48HVS+u0ElE5Ntqgy1TjbZxZad6JRkI2aAeGl0ss\nKAeHSlqTU56obl5xtSTkiKzaULMQJ7C5dSLLly/nvXv3xtrHNaJIJlpeQ6cqx4lWSqvkuG6Mca/x\nodVLvTbhaRu6nCNgctWty22KBnWonn/xfgDAmm3DxrHIxxscKhlL9MjIGkbUjHvdlmed3nlbSK1q\ne9fkxfqx0j5mXm7b7qLWLOKW/I6ajkyqsuuxm1Uf80RY8YFZOPJOWSm0kpQ1NznMBodKdaHG3p/h\nSRudD5Ewab5RCQqVucmkpTy37npjyR2gvp3yvdv3u0dCMnBEUzHB9Z2PY+0Q22fpu7iohYXril4l\nsW2qsusDIcJxkybejDPjpaOntSuKZjr72UINgWpOysDj9nBDj8cVlQ9RZ7YR6KL0bJVjRckdXWTT\n5b0F5wjE6H46emO0OHBJ2pPJ0ndxUffgNiW7zOjOG9somsJtr9vyrNMKRKyE1t+0SNtwyCUfp1wZ\nx73b9yubsqTV2Q8InYoK30plnH0rVU9qqBoJmd6naGa2jEtXuf6+Ir722SUN72CQp1h+g+h4dMSx\n/MdtmJRlsuxFrVmYVgpnKxPawnuDQyWtzRQwayymfAWV/0O1mtddi8p8lDTSK/rQDQ6VjNfly394\n0iSac2Sy84vFkthPJlZlhOgrFv4e59lWmWWjvk3XYBVZCPb3Fa2+jqyTZS9qYWFyLusEQRKVVBAt\nOSAf01R2HJgUJCYhVa6MY8POgw3H2nzrYq2aHeSAsYnG2G75oRPXbMKX//BkiW3Ro1ssuVZGUGnN\nlYmqxuzSJM30bkezxHXM6gnQ062PsjSZ51qRLHvRR0PpoiHyRDi8+dMNn+uku6sjKur/MEVruJRD\ndz0f4JZAJHJGHuifLPprW9EEecLW25d4n4UnU1zK3JhK3JjQRWARgAdXL7Vq97rzLt34jJMmoZoX\ndJaGtEt9uEZDXdQ+i/6+Iq67erbyb3ddO1f5uU4lnYgZsSAwOcpVCHuua3Ex2Vm9+dbF6LXUxld1\n5jKp4bN6Ai8oPKkjmoldtW4XrtvybC2HQOVbkElqDrX6Niyvm+68JkEh+2RUC8jSSBmMyUhLlU+y\nlVzUZigAeOQ//mvcP3gAj75wrG7F8ugLx/DwnqM19Q6oTuw6kTCzEOD02YqT88qlz69wlJtMU64a\nhjiHrmiibnuxijHFvPsEPk/a6MLS9755Ertefrvp1qsqVi6c0xBCK0ptuNSA6u0JYnXKBNCQD2Ir\n6bHxyYM4W5loS10owAsLAKiZXFQ9KmqVKElf/jjIEc6cH3OOcnDp8yvHlKs676nKq+vq+Mvnc1l5\nyeGCOmHkK896skKnbdtaA8jPZBxzzeBQCTv2leqOTQA+Mm8mduwr2Yt25gnvna0vCigmcV2WeE+Q\nqwmH3p4A750dq5m5dOdTHadVdaGAi9wMJbD1qKhMsFZQFHsLuGR6l3G1IxOdZAdWLUAhyNdto4sp\nF533ZHX03NhkKWRTwxeBbeUltjeFC+rCiT2eNNAtaGxvWLRqgYsZR/hBos86A3ju8Emr5j6rJ8DY\nODf4M8QkrgqLz1F1ThHjOzVasUY7mmhVJKIXFjCbl0wQqqqkqWgfMJnPoZpk48aUR/tsmB5m1fl0\nwim6ve4BFNcMoMGm7PGkQRJTUrG3YKxaoPIDNhPZCFQnz/fO6fvbvDVSrpp+b19S937PLATOi0sX\nWhWJ6IUFmneK2b6sceZago9qNd7fV8Rz667HG1tuxHPrrrcm4IjxJhm3LJyAqiBThd6ZHH6qldua\nbcPo2/SMFxqeplEtaExEtWeTH1Be3CQphSMzAb1pGph8h/r7ihhYtQCXh3Wt0m6T2ipzsBcWaM4p\nBrg93JVxxpptw06rcNvxXIWUUL/vHzxQpwXI54hWwRVjU43BZqIS7Wi9wPDEIRr5BKAW7m2jtxA0\naM+m90I2S6VdwFMmyBPOnBvDVet2YenGZzDw+P7a4ipNZnTnW2YO9sIC8Vcygh37SrWQPtmUZEJl\nP9W9LLN6GsNc5VXUwKoF1vMJx2DUfrvxyYNGVV1lHrOZqKLH8HhsqLTUtduGsffNk0YNu9hbwEOr\nl2J4/SeVWdu29zmORpGkB7ZoK8Co+hqTmJ0IVUe4iSDfuincR0NhMuws2nvChq5Pt23FIkcw6MIE\nN9+6GENf/aQ1u9tWNRNodAyWK+Pal0UWBLo2r7Zr9KU/PCpUz7JKS2VUIxPvXjGvrvEP4FaGO5q1\n3cxqPm5SazPkUDVtCRhmMxcAnI7Z56YZLuoM7iiudeajxK07D1RXDW9suVF7zjg5DHInvmZxOa8t\nrNbnX7SWpFm9zWYDxw1PVVW5egnLAAAgAElEQVQqML03oiqCqrtkHJK+10B9rSf5etOYNaPlPZJU\nn07jXfP9LBKQZEWcJ0q02hB2VVsJZRdUORc2egsBzo1NNLy8Ls4ycb4vP/FyrZ2kjPDltAJby9q4\nx0mrjEIWZRl054lqpnIbXt25XVqXmq7DdX+BLkLJVCZHfC7KdNuEke5+N9M2WfjhxHXFbWCkI8gT\nmFE33rWWRkxRCKjrz511G2SvWUg0+wBE6S0Eysk7yBG23rHEWEmytxBgeP0nnc7TTL2oZia0pFpR\nkolUtQ+gNw3E6RoWtz5Xq49nwuWZVZ3b5bszXYdJk1UJa1P3O1uvCtXYZFTjlAvsrVw4B7tefrup\nKCSXDnwuENCQhAdU7+v0IBdrjB/86Rk4fups08+Zrw2VgLiO7hxB6YQGqg/XjGlqxe2S6V21L1O3\nCn/33JhzVNGGnY3Oah2yozoasht3IkuiFSWpe6PbR+WkF9gc7XJQgSopqxlHfdx6X83gooGqzu3y\n3Zmuw3Re1Xeqi1Aq9hZw94p51kAN05h1fg8xlof3HG06XDV67mgAyKyeAEHOfBXF3gLe2HIjerq7\nlEl8cdftr58YbdlzBjgKCyL6GSL6MyL6bvj7h4joC5mMqI2IB8CVCQbeOzvWkKEpVGZtcTHpwY0W\n7ROMT7g1FBocKjlXtXxo9dJEQkGHS2OZKEkmUt0+tglAd/+jwkdnBknqqE/DtOiKa9h39Ny6/XJE\ntUnedB2280a/U1Mo9gP9i/Hg6qW1iVcXfaQ7ZysCKlTnlvMnRkYruGR6V61QZ/QKRPkekyZ4ulzR\nLj5VpP3c2nDVLP4CwNMALg9//xGANVkMqN309xWdu1IB1bT9sXHGrJ6gIcTUZTI1fbEuX/rGJw9q\n/5YnMnb7c0FV/VNgmgB0JJlI064k6pqMlTT/JokQTYqrNhw9t24/0RdicKhkvA6X80Yj63Sh2FF+\nanqXdgHmcm1JmNGtvxZ5oo+GvMuLjlOjFZw5P4beQgDGpNCTzWylkbJWi7q8t4D1Ny1KFMYfPU4W\nuDq438/M24noPgBg5jEiyi6GrM3EdYgxqp31Hly91NqsRFWrSbfSsH3pg0Ml4+r6a59d4uSsNB3f\n5MR0bSwTvSbV9YprVY1Rt4/KSS8wTS4uwqeZQom6JjViwknTCRn9DlT2cAB4+3QZ9w8ewAP9i63Z\ny0IrMD2/8nldn19dKHb0ORspVxDkCLN6AoyM1jvpVc/HwKoF2r40LtyzYh6WXzlb+87LE738/Kvu\nocivACZX/qoGfFE/TfS+xg3jF4hKEVngKizOENGlCK+PiFYAOJ3JiDqA6AtImmquMqrqjy6T6cqF\nc/DwnqMNx3P50k2mm1k9QYOgiBO9Io6vMxnJuR5xOv+ZJqDBoRIGHt9fiy0XkT2rPzq3Id4+yBGI\n6iNqVNFQ0bHMv1RfeytPhAnmpqNKohNpdGUpks7kBlPNEP0Oqt/1yyhX6otMPrznKLb94CgUAWwN\niLpG4jpsi4Ho5CcidVxQTroTjJ7uLgx9dTLIQxf5NaM7eVDnPZFGX/K1njk31mDilZ//Zsw9wvmu\ny5/q7yvi7j/5Bzx3+GSs487o7kptIRLF9S5/CcBOAFcT0XMA5gC4I5MRdQjyC+ja7Ur18OhWU8Bk\naeQoBGD1R+dav3TTw7r+pkV1v7tM/K7Hd3Fg173Qj+3HxicP1laJty0rYvehEw0vSt+mZxqSkCrj\njF0vv42PzJtZ9+KMTXBt5SVCK6NmDdVYdKvgNKKVdEln0XOKpLPlV87O5MXu7yvWelJHcREUQH1d\nI7kpz9anX8XabcOYWQhw5vxY7ftSrZ537Cs5XaOplpOMbSUfB9ERcvehE7hq3S5cHkZNyeiO+9ZI\nGYNDJWOLYxsuEYMvHY2/Hs8ySc9VWBwE8MsAFqB6n1/FRRRJ5fowxrUV6kwBri+athcGAWu3DddM\nCYC+96/qc1vTo7gO7Io0sZdGytixr6ScmHVq96nRSsMKS5WVHhV8cQrFpSEoVJqb7vwM4N7t+7F2\n23Am8fFJJzFAbYJTmYpsuPZaMJli56/bBUAfhm6CCNoII/GOyd+XrOFHNUKZmYWgqWq1gD0/wvTs\n6npkANlWoHUVFv/AzB9BVWgAAIjoJQAfyWRUHcTgUMk5Drw0UsbSjc8oG7frttfh8qLpfCviGS6N\nlHHvY/uNUj0aedJs06M4oZxRM1GzyMcYHCrFyplxnajlrop5Itx17Vw80L+4qaSzOB3PotrLyoVz\nlFqaa094GQK0gitphVaX71UkpJlGm0R7sF2+7Xp0u58Oaz5FyVH1fbL1pghyjYIq+v2b7ptOUBCy\nrUBrFBZE9LMAigAKRNSHyYiw9wHoyWxUHUTcXhcj5Uq1sx4aX3z5Re91CJGzvWguTsbxCYbplYhO\nKKZJobcQWAWhS20sYFKVH3hsf1ONX6LnBiYFniuuheLuHzxQt/ocZ679rvuuxpmdFhvlyjjWSNqg\n6h7fP3igrg6YajUsrvuua+cqfWEm3thyo/ZvSYV5b09gDazo7ys6O6hdF25Zojv/BAPvm95lFWxj\nDFQsJmHX98hlXGlhMyWtAvBfAVwB4OsAvhb++xKAL2c7tNaiCxFN8pJUFDkSqjA7GzMLdoEiEuvi\nhPvKRPczXe/pcsVaZj1OKOeGnQcdVmEES64TgHonuSrJzsRd18512u7RF45pP08j6QzQJynaujkK\nxKTzQP9i3LNintO9E+M0kdS8cXq0UleeO3p94r1zRTiGCeZw1yhJKsfGheCmAek0HvndS1oJO8sW\nAUbNgpn/EsBfEtFtzLwj7sGJ6AYA3wCQB/CnzLwl8vd/C+AhAB8GcCczPy797fMA7g9/fSAcSyaY\nIoWSSHigcdJNosYTuYe7JhFqKpOS6Xp1IYTRMcoO7KgjVDB6vjHSJIoo4rb3zZPKVXJPkEO5MtFQ\n/iOO+aUQ5JyjkkzmJFuY6fIrZzuvnnX+F9erKo2Ua7Z+lylS7r2gM22pri/IES6Z3mVc+EwAmIh8\n9+XKOO7dvh973zzZEOVmQ0S6xamDJqKOsiSOxqMzETLQEFodt3Bhlj25nWtDEdGNABYBmC4+Y+ZN\nhu3zqCbvfQLAcQAvAriLmX8obTMfVZPWbwHYKYQFEc0GsBfAclTv4T4Ay5j5lO58zdSGMtXJ0cXM\n2+5aNNrBVBvHRLQyp4jiiE5wcetaCZMSgAYbuOsLrLs/0ciiwaFSrJdbPr64hzpfgUzceyDX6HLh\n6vue0goMQlUTJEJDbkCS8YmqxIKkz4+NnENYuHjmll85W+kvSVpPLYlJaUZ3HufHJmKZLq+7ejaO\nvKOPhEuKHGrteuxCkMdty4rGdyxOLS8V0WfHun2ataGI6I8ArAbwn8Kx3AHgSstu1wB4jZlfZ+bz\nAL4N4BZ5A2Y+wswvo76MO1A1f32PmU+GAuJ7AG5wGWsSTCGiqsxTm1khyDXmSCRR41UVbUXYZVTV\nHFi1QFmbJkfVWHJ5/KJpDICGmks79pVw27Kik9r+1kjZqXxHf19RWyfLdnzBA/2LcXjzp3Fky404\nvPnTSm0gzuqxtxAYBcXgUAl9m57B/HW7MD/sdrbiA7O0xxNNbkRypiirIps3z5xrLA2jY2YhqDOL\nupgk40BUfU5d5lzxzAHVKJ4cUc1f0swEnET4nTk/HtvH9fzhk1i5cE7TmdEyhKr50tYKOUfVZ03O\nWn+gf3Fda+MoqvI3KrOUKRM8C1zf4F9k5g8T0cvMvJGIvgbgCcs+RQCykfc4gGsdz6fat+GtJqIv\nAvgiAMybN8/x0I3Ysop1uRI6G3JlgvGl7cM1s0NvIcBnllzWmFiWJ8zorjrEVBmdprBLXQKgvIKX\na/Gr0E30j75wDHdd25gIpxqHbrKITtxJzADyQ+9ijnNd4bnEuMvJgUBVEPzgjVO47urZ2PP6Ka2G\nYWpsJTKTZ3Tncea8WXMbKVdq32NppIwgTwhy9ZE2BOAXw1Vz3B4LzEAlhrmOAXzlOwes444S5Agg\nexOfKEJrTaNPCwM1rTQp07pyODc2uaaNhrcPrFqgDNaYYODcWGN1B/GzzjSpKlwI2K0AzVQesOEq\nLM6G/x8lossBvAPgKss+KsHn+m057cvM3wTwTaBqhnI8dh2DQyWMnh9r+Nx20x/oX1ynlkdt8/Iz\nM1KuYNsPjmH1NXOVYY5iHC4JXYK4CYCuxwCqNnihYSQ1M0RXN3F9P/L9d80+12XDRzkTVvTVlZ64\nd/t+5cRSmWD88O138bMzpxsnZ3FfdTknM/I5BLl4ppTKOGNaVw7F903XCsy0S+xHiSsoxIQPQHtP\nVYsk2QQTt8eDDpOg0JnD5M9lQSFo8A9olvoqP4ItYk9XuDD6zEZNg1n2s3AVFk8SUS+ArQBeQvUe\n/olln+MA5DCTKwC85Xi+4wA+Ftn3+477OqPLKXAJEQUam6GYbPKVCcbuQyfqegVsffpVrNk2XFem\nIroC0cWfN6Nq2pLugOoD/vCeozWzVdxaNcJh2tsTgDlenHyeqDZh6CZv1Quoq+AbTc4aKVeUJTfE\n82CaWE6NVqz3wdbYKknOAFCdsFYunKN1yA+sWtCgEbULWTPQBToI+71uAQUkDzCxIfsbVCt0V3+K\nvDAw3fc4AS8qM7aOuAvEZrAKCyLKAfh7Zh4BsIOI/hbAdGa25aK/COCDRHQVgBKAOwF8znFcTwP4\nPSISRuJPArjPcV9ndF+YsK/H6UDlYmYRheSiD6ecnDXw+GSORn9fURkJFOdhihK3aYstE1mHmBCT\nFEObYK4z47iWYtZ9B6rdGY21klycvS6URsq4+r6nMnFIP/rCsVoxQJXJcevtS/Cl7cOJr8OU9exK\nIchj5cI5zsUBTaxcOMcpZDguE8x1TmCxQpc7LrogIhZt7390cWfcPvso30RYHdzMPIFqboX4/ZyD\noAAzjwH4DVQn/lcAbGfmg0S0iYhuBgAi+igRHUfVYf7HRHQw3PckgN9BVeC8CGBT+FmqmGrSxG3Q\n47rSL42U8cieo9rJtzLOdWXHl185u9Ep2sTDlCSEN0nmbjOIe2kbq8rUFRe5VlJKuYEAmiu3YTuu\nSGaUNZRTYT4DAHz9s0sbnpkcoeGzfI5qjxKFvycdtgiIEE7c3YdOaIsDujbbErXTdENqxmGtqogr\nnMhxvrsJBgYe328MQFAVVTQ9q5Vxt142rca1vtMzRHQbUbzMFmZ+ipl/npmvZubfDT/7KjPvDH9+\nkZmvYOYZzHwpMy+S9v0WM/9c+O/P45zXFd0XpopCsjXoiZNEY3sU5dX4xicPKovruTZGiiYatqJR\nTDMEOcLo+aoJy2Z+mH9poe76VBEvhSDvnJg2FcgTVU0eCslW91xw436rPzq3rrNbTtqMUc32T0Kx\nt1CLVBNCIE4tMh2mxQIBuG1ZEYUgfok6nT8yaUmTyjiDCNrnTDjDbb1gZDrxPXW9018C8BiAc0T0\nEyJ6l4h+kuG4WoKueU+SDlRyiC2gf3DiYOpXYXuYVK1I124bxvQELxfglgGbSpYsVYWly7T1/OGT\nyrBf8R0IoZ+mxtBu7rp2rnHCLY2Uce/2xqicysRkaZIHVy9VtvZMiupZ1D0Lqs+TVE9gVH1UZ13L\n6EroikaazmfLcD81WjE+Z6pw8tuWmcvmRDE1ImsFTjMHM/8UM+eYuZuZ3xf+/r6sB5c14gsTD3Ce\nCB+ZNzN2W0f5eM+tux5HttyI1zffiCNbbtQ+ZKZptbcQ1By7OqJx+NEHR9eXuFyZsPYKjlII8rjr\n2rlWzWmiSdNLniiWcza6Zbkyjt2HTiQyJziNr80qiui9YBPKpuuWTaxpoXovbMUTBaae7Lb3zdTe\nVXePir0FrfnLVLKlmZI6gqgw0gZkoLEgoOo+rdk2jKUbn2mZ0HBNyvt7l8+mGsImKh7gcWY8d/ik\n8kFPGr+s017uXjFP2W83yBE+s+Qya1TOmfNjRp+KaZV0yfSumjnChlD3RSKRiWaTgdKY3HWJgnER\nUWByMuPX7liCnoSaWTOI/ukiCqrZ+ySq4doI8mR9RnTvhW5ijX5uSuq0mWqEkzy6+AlypFzc2N5h\nW5tgXeKrKzkiXLVuF/o2PYOlG58xltVxrforIvvuH3QvnJkU45NPRNPD0hvvJ6JZRDQ7/Dcfk/24\npyyuk4ocymlCpSaqMsBFFufQVz/ZMCFtvWOJ0jkoQ4okp6iaa5q4R0YreG7d9XjDoPkIZHtrf5++\nP7lYDWVRsG1WT+DsD7o8hTpAck0ncZ8GVi3AxicPYjSB2cOVGd159IaO0qjDWH72ml3hAmqBI6KV\nas/i7UuMx1CNTbwDql7TqsnapXpCr8J5XHes6ImoGhji2vNboHtX5Q52W+9YohyPC+PMtSKipvBp\n1fdrM8mpqjqkjbE2FBH9JoA1qAoGOUfiJwD+hJn/W6aji0GS2lBx6u3IrTpVqEJSk3ZfM43LlNkt\n14T5xNe/jx//8xnldrN6AvR0dxkL/UURqrjqOuV6VdEy3s0iYvG/81LJmhQm7nfcrF9REG9ktKKs\n7wQgUfiwK72FoFZ+xYW44c8uqGqOmRIUVVnwumeDoX9/THXZ5OPrMvhd90+bOAmQcUJxdXNG36Zn\nrGHoSa85ldpQzPwNZr4KwG8x81XSvyWdJCiSEsdsInIgdNLbpUZSs+MSGo5uZSPC9waHSlpBAQDv\nnZ00YY2UKwDbHfJiZaNafT0omUhEaew09AvhoH5kz1Gn7GHhvI9b3vmS6V1Yf9MiPLh6Kc6NTdQc\n7MIuvHb7cKbhwyPlCvo2udueha8tTR1O5J2IcZhyXOJEEzGq36MwD6pqmrmYi2QtTw67TdL6Nw1s\nx88T4Z4V83Bky43Ovjyd9jM4VMJ7ZxurTMQdU7PYmh9dz8zPAigR0a3RvzOzrT5UR6PrNKdD5EDE\niaTQfS5WSnISkFh9mSq5AsBPzmo6ZYWzh01AqSJlCkEOZysT1laqqk5toi+zWPE90F81s+l6lwc5\nt17QYqJy1f5OjVZw3xMHsPnWxbhtWdFZwxH7TevKqdvctiCaSowBMHfLk5+brMZhKqUeNcnKz4Pu\nNpk6AqpqHqk0EJ1mYavrlhRbLTJtS2OgoQqDSxa6rBVEz33m3JhT5FqWLVUBu4P7l8P/3xT++0z4\nT/w8pYmGu7pwarSiXAHqvihVso4c2QCoXyaV7RSomkN0z81IqKYmWWGUDYJCbiwUjcgQlUdVjvbP\nLLlMeTxFmZ3UENqcLtLEtF/SMhxpYdNEo8+NCvG8ZIXIro+Ox1Weqq5RpzUITBFTKxfOcfKNxMF0\nPoGuCuzdK+Y1jN+m6arqoMnndnkus26pCtjNUOvDH/9PAH8H4DCANwEcCf9NeZJ0mlNlc+siJUbK\nlYbwNpNjXa55NLBqQc1hu/XpV7HxyYNOGc1prjBk1dglIECeDHQTdtYL9dJI+r0LWoVJ0Lvcf7EK\nTrMct4zcDjjLvtwyOhPvhp0HGzK8RfRekmAU2/mieRImc6xMdNtZPUFD2XJZw4rdJA1qIZU2roUE\nBwGMoFpEUFSgvYBSneKZpFRF7Pr7itpieyPlehOD7WUR/amjlVZNREP80nCAElDnMHN9ycV2ifs2\nF4K2rPJn9QSJalmpjhO3eKJAZ+4bWLXAej/lKC4ADaaMNO7pqdHq4mfDzYsSf7+6xYzO9BOnIKNI\n1jNhq2LsalKOU8TPddu491TXDC0LXIPGr2DmO5n5D5j5a+G/r2c6shZjCtNTofpSRwwTjWgledW6\nXchZwksv7y1YtQiZqB1ZtepR5XQIdM7tpLWXbBqOKVeh2FuodfBrJYUgj/U3LUocFinT092FDTcv\nir26N5n71miqDwtUYZ6yaed0isJXLH56Dc+UDsJkQU15Nd9Mcl4U24Rr0xx058vaJ2A6x6yeQBmW\n7iIc08JVWDxPRNmLrjbT31fE8PrJ3AcTqi/V9jCJOGtbGN3KhXOcV7iFII+vfbax41t0sjAJsglu\nLMqmsvu6mDeiGo5qe12JMbHv3jdTrxmpJWoKSDLJR1F1WLTln8gCP64porcQWAvzpT3RlSvjOFsZ\nj32vxJMf9QPETc4rBHntAsgl89v0uWuEVhbozr3+pkXaqKpW1ZFyFRb/BsA+InqViF4mogNE9HKW\nA2snctmOh1YvdX5w0rIVP/rCMe3fegtBrEQjgekFEsdROdRluy7Q6HiPtmyNrm5VIZ66MFhhazZd\nv4m4CYHF3kKDUzU6ySfJ2BaZumKie8MhfHKcGWu3DSdqYDRSrtTCXXV2+Cz8GOXKhLG+kX3/ydV8\n3NbGm29djPU3NQp2l0ndpjnYkvOyxHTudmo8gCUpr7YR0ZWqz5n5zdRHlJAkSXlRdDZTUxid/LeZ\nhQCV8YnYHcXi8FAkLM8VUdY6GoIX5Albb2/UTJpNMkwS4imS0+av2+W8TxTXpjWu19Js9zmRVBin\nrafrNcgEeQIYDW1X5YQ4QN/GM+lYhAae9B6JUNM4yX8yLu12VfuklUDbSnQJr/c06bNwTcpzEhZT\ngWaFRZIHyJZJK140XQanyKR2fdHiZvqqxuvao1s3SfYWAsyY1mWNiU/qYD+y5UZcfd9TieofuUyy\nBDhPKkC8LP9mxpU1tmdRxz0r5mHXy29rzaLCwZq0QdGsngBnKxPKZyXLCTyJkGk3urylZucFV2Hh\nGg11wWOymeoeIpttWV7VqQTR+puqjlyX1V4hyFsdv1EtR1W2Ysa0LpwuV2r5H2u3DdfMJfJ1miJQ\nxAOr64XdTCG/waESursI5Ur8qce2R5KXKo22nu0WFMDkGOIIChH4cN6QGFMIcsYGRSYKQR7M6uZa\ntnpszU72rWxHGsXVgrFy4Zy6lrO6aLZWRQ56YRGSpGyAi2NJ2FwBdZaq8AWokPsE60xfuvpF8gNU\na9cqmSmif49O+q6TpEqgJp1cZ3Tnw2vIJmvvzPmxOoeqy0STVVvPOMTVBtJigmHNgm+msOLmWxdj\nrWahJCf/RbGFvrpgW1hlJUh0Y9/75sm6Vssi4VVge6dEoc8s8cIiRDc5MqomGdUD5DKhyk6zuA1X\nVFFOuodNV6pCYCsUGJ304+RqRK8hyeQW5AlBPpfpKqkyztiw8yDOjU04TTS6tp4zuvMYPT9eTZg8\nXc68HMhEWArG9KypfBadjoj8iluuwxb66lI6xLSwiit4VOi0B93Y4/i0VJgsIGnR+uL8HUq0R66M\nrgd3nDR+Haa4atWXr3vY0phk5Uk/TimU6DXEeejlcthp5gLoGClXnAs+6sxpvT3deHD1UgDN142a\n1RPUIsp0EJlXluL+bb1jSe04nd5JVowzSZiqboElN3bSlekA7GbSpAVABaacEd3Ym9UcWxE+6zWL\nEFtiiy5rG0AsdVZll5TVT6DenxEly4cimmQltCFTRJDqxbatguXt5EiXLAvk2VCd1zYpJfXL6Hwn\numgXk7IQvYdRU6XoK5GlvhE9fiHI4yPzZuL5wyeV55WfGTFeOfDC1vrXpNG7+B1dzccqTAVATSU7\nxDh0Y2/W1NiK8FkvLEKSPkBxHGUqE5LoGy07skw2U93DZooqAarOSpuFQn5WozbdIK9ueap6sV1M\nWEGO6oTM4FAJo+ftZZizQpWjEWdSisPpMC8iaqaIm4mrW4HLxy6GC5JmzRwm5KPKE6fLxCo4JznR\nbRV4B1YtwFpLRrtM9L2NYz6Wib6/umq6Jv/ng6uXKoNdbltWbMo31oqEQS8sQpI+QHHQrTh2Hzqh\nbCSjsnkOrFqAgcf3103cQZ5qmohKyxENjiYsfgthBlLZdEUXtVOjlbqVpOrF7u8rYu+bJ40Pf2WC\n61T9pCv1PBGmdVHTHexUE2laNbaizCwESr9TnPOoooWiodHi2Dv2lXDXtXMbNFgVQZ4wo7saMdfd\nlaubxF1YuXBO3XPgspCKG4nY31eMlS8SfW9t32ucfh2q8ZrKppuCXUzBBMXeAkbPjylDmHsLapN1\n2nhhEZL0AYqDa8SVKdoDQKNNIfxd93Jet+VZJ5+GeKlUL4VwnKpMTKoXe9fLb1tXSa7OeR2iw10a\nxf8AYP66XchTtX/zA/31lUDTMo8Rqj4I1eQYxxQRjRYy5baIBYnoIigvJk6NVrSr/iQJiY++cCx2\ngliSSERXU6eqdHcS87FtPPLfVXOJXBNrYNUCZaKh7ppMXSpdQurTwguLkKQPUBx6NVVNo74CW7SH\nqnmRKRrCxcQmC0Pd9qdGK9qJWX7I7x884DyBlyvjzoJCTggU2lJagkIwzlxb4QmB0d9XTJScF22B\nKyew6c6tM/dFuby3UKd95iyCRnw/ulalKvNQEv9YXFPX4FBJO3aTJq/SsFUwGv04Lk2W5GZeruHk\ncuQjAKXPyBRtpcvHivp32pVM6IVFSLNJPqZEG13pckH0PUk758P2kMvlIK7b8mxiu6mIOtFNhs0Q\n5AhEqO8e5jCpJiW6Qk6SnHd2bPKl7y0E2HBzNVt+96ETWienyzUVgjxWLpyjtJ+bWLNtuGbrz4Uz\nmDAyRScx0yRuwzXmP0nr1jochiaKDbrkZti2MVkfopqDKThEZ2JzEQbtTCb0wgLNJ/mYEm22vXjM\nOgHIIaMuK624cekDqxZobbyiZ0UzJToE9z1xANODXGxhM6M7j4lIJq8wMY2MVjA9yKFcmagJ3GZM\nQgS9hiczzlz34pvuoQ75K5Rt/ysXzlHap3UTcw5AoTtfqzk2rSuHXS+/nei7EmdQBTvI2uvAY+pa\nTS4MPLYfX37i5ZofSZSVAeonwtHzY4myt8VxXHJK3js7VlvI2fwitm2iGoMw35k0B5eF31QpPeKF\nBZKV+nDZ3zUCRW5447LSMqmqKvr79I2Zerrz2mtQYbKrxzEpyQT5HDbcvEirmemyfJPAANbftMhJ\nMMpCv9meAaKfydptw9Z+JlEmAIxKxSmzTFwsjZSxYefBppL7KhNct/+p0QrufWw/cpg0oZoEvil7\nW+BqIhMmWpdJ22Wb6En1mFEAACAASURBVMrepjnYeoTfP3igLhAkraTALPDCAvaHxFbLRffguwgK\neaLXTdiqlVbclcj6mxbh3sf2YzwyCZw5P477Bw84vXxBjtDdlUu9qu7pckWrXm99+tVUcwTyRLXz\nuGgK5cp4auU+xPOQZMUed4+kuRWEbITR+ATD9alxiTqMYxYU74mtQoNtYtcd2/S5yQ8xOFRSPlvy\nwkJVI6pdmofP4Ia5vr0uG/PuP/kHrN02bHxgbf0VonXydQ9edKXV31ff2AiAto+BvM+EZrX46AvH\ntPcgT1Q13RQCgPS9KAS9hSB234SZhUA7/rSTEMVEHedl00260W+3k7KmhcYYl3YXC3GNOowTmWjr\nSy7e6ZUL58TOJm+mN4ZpISQapYkaUbas9FbgNQuYpb/OxPTcYXM3N5FoY4qdjobPFYKcMl9gpqHV\nZxx/i+nB1N0D8WC7hN/KYXw6s1cOk05VoKqtnDk/VlfJdu22YazZNoxib8HJvxAHuaxGsz23VRHM\nrSxHbupVnmVPlTQpBDnMnjEt9qrZNddCJH/awqBV4cW6Ap6yv6JXkbAaFTBJ6sKZiGMiTxMvLGCO\nQkhiL5fDEHW9AKK1gAaHStrEMpOCEsffovM3yKYZ3YtierBVPSL6+4oN9lgAyOcJ7wuTvoSTM3p/\nZPttkCPncFIXSiNlzF+3C4Ug5+Q7iDv5py0oegsBPrPkMmVJGOHnaWWJlLSF4fQgb2xuZMIp10L6\nim1h0KrwYoEue1tOWI0bZt9M+ftWtVKV8cIiRCf9436h0Vo9KmeqSrU1FS4z9c+OE2Z717VzlZrO\nXdfOBWAOy9PdB1Mns92HTjS8lJVxxoxpXbXaSLaueJUJNgrLpJjKoMsd5lS1uwC38iku9ErJcUB1\npT09yCsnHhEwkSeqtaAFkme/x0UsguQFRbOCSjzbSbpU6qLKZCrjjTlIpnHrtHJTAEhlgtHT3YWh\nr8brldJMhYBWtVKV8T4LCypbp27u0mWLuvTzNb100aQ9mTh9eR/oX4x7Vsyr+VLyRM4tGZNUB9Vd\nk4hHd22f2upWDsKcNHp+DI/sOYrpQQ7RElhpCIpCkMdnllyGs5LgEiHCwj498Ph+fOi3v4uH9xyt\nc5Dv2Feq5TNsvnVx1aeUEAr/mXxsIrcjWgQzbt/zKCa/4P2DB4xVZF0j1KILJ5P/QlVxdnCoZBWK\nSVb6qrnBhTSqSSTBaxYWVOYZ1WpTZOfK9s1oC9MHNf2zB4dKRvU+OlnKq63engBBjupCFU0P0wP9\nixP1602SPWoKs21XdVlXGJOr/TR9JtGGVraQ5co4K01w5co4Nj55sC6D28Y0Ta0nBtCdJ5w3mPpu\nW1Y0NudJghBAqv7butBz2cTqOkFHF062aLi3RsoNVXvjnsMVl1BcQN8IrZV4YeGAyjyz/MrZRkfY\nwGP7G2LNBx7fXzuejC08NJq0J6uup0YrCPKE3kJQ8wNk9TDFzR7Nsrubqx+jEORB4KYLDabFODOO\nbLmx9nszOSRy+RWXe20qCmgSFL2FALsPnUjV1DWrJ8CNH676YnRj130uhISLGUy3cBLRSMo+8z31\nhR5tdzbNlb4t0KSdeGGRENPEqcsuVdlPAbsKK69alEX+In6ALImTbZplO9AZ3V149+yY8fiEagn1\ntOtHNUPUbJOG3T9rRJmVNDlbmcDf7jdnoeueH/E+qCZWuWqu7fnUTcy6vuAqdCXXk9Lu+k8mMhUW\nRHQDgG8AyAP4U2beEvn7NAB/BWAZgHcArGbmI0Q0H8ArAITxcA8z/1qWY02TuDWcTBNGdNWSpG5U\nlKTlBVzDdMXxs9QsXLrqyaakTiF6T5KUEWk1p0YrTYcZR7Fl+wc5wuprGsuqp1lYT7e/i7ZnCuxo\nlnbWfzKRmbAgojyAPwTwCQDHAbxIRDuZ+YfSZl8AcIqZf46I7gTw+wBWh387zMxLsxpflpgmf5Vt\nUxcVIWrquERyuNpMm6mD5RKmOzhUcqoG2iwzC4FVs+hEhBOzvmJsOg5zILs8j/fOjqUawmyFqqZe\nk7kXSDax2hZLtnDkdjmY202WmsU1AF5j5tcBgIi+DeAWALKwuAXAhvDnxwH8N6IsAiWbI+5KfGDV\nggafBVBVkXX2U6CxQFlPd+PXYytjbKOZOlguWs3GJw+2ZEJ591xnC4resIS6KlkrKrDTvIys7oh4\nlpOYFnUCrCfIgUFKDUOYbJ9bd32qq2yXxZKuH4UIp3Zpl9wppqM0yTJ0tgjgmPT78fAz5TbMPAbg\nNIBLw79dRURDRPQ/iOiXVCcgoi8S0V4i2nviRHOF3nSYmq/r6O8rYusdS+rCGWf1BNh6+xIA6tIc\norJpIcg3tGuUz+UaiqujGTOWS5huq8w+0RpXncRDq5dieP0nsfX2JdoyDy428axWTc0cV1RZjYPu\nm5oWOm51ZJF4plss3bt9f+2dBNDwjj24eimOhOV1VIIiOkes2TaMvk3PtKUsR1ZkqVmonqnoc6Pb\n5m0A85j5HSJaBmCQiBYx80/qNmT+JoBvAsDy5cszmT2SrsRV6rFtVeN6rmZsms2YsZrVajqBnrDc\nuQiBTrM3dTR8Wvc92eqJjTOn7iOQjy86AcrhobIpjMis7aT1oo2MVoxRSVkknukEUHSBtvnWxc4+\nCZ3wt/USB6aWRpKlZnEcwFzp9ysAvKXbhoi6AMwEcJKZzzHzOwDAzPsAHAbw8xmOVUuzDuXBoVJN\nk7h3+35jB7w0nNc2dMl1KxfOcSpGaNNqmkkOawUMwoOrl+K5ddfjgf7FmGhCUPQEudr15onAqCaK\n2VaTJkOr6Jb33tmxxOMyIRL67h88UJukZvUEdVFarbLuyVFNcRM+mz2nCVVingnT+2k6VhKrRTvJ\nUrN4EcAHiegqACUAdwL4XGSbnQA+D+AfANwO4FlmZiKag6rQGCeiDwD4IIDXMxyrlmZW4rp6MlFK\nI2VctW6X1rhrO1ec1YlLkqHJ6W3TajbcvEjpr+kUoppa0tDVe1bMw9/uf7uWdBldmYoeGKrvxDYZ\n23w+zTqxo2XX2xExJvvvWhku6lpiI84CzfYM6Y5la58sF+OUOy22i8yEBTOPEdFvAHga1dDZbzHz\nQSLaBGAvM+8E8GcA/pqIXgNwElWBAgD/FsAmIhoDMA7g15jZXOY1I2z16G1RFa7x2lz7Tz2iaqaO\nJNFNqqzRZpo/RY8NuPWKiJJkEiQgdvtPIZx12fgubPvBMa1AjE7GaTe0YTT2905yjHYSFYhZhotG\n39PblhVrgjxJ/+8othpVumPphIgo8yLfo5FyBQOPqZN6W0WmeRbM/BSApyKffVX6+SyAOxT77QCw\nI8uxqTBN/tHPAVgn6TTMR5dM71LmMMi9qJud6JOYv0z3qr+viLXbh+ObMwgIiJy1khndeRzcdEOi\nlrBC7d/24jF05eK7fG1j1DW0SQMRkaMqkwG0tkx6M3z5iZczn/hUi6kd+0o186nq2YlrAjPVqDId\nS6eR6HqxVyYYG3YevDCFxVTCtkKPfkEuq/E0q3LqxqjD1uVPJq6pzUWb+cUPzLb2/IjCDCBn7tMg\nc+b8eK2YHgB85TsHYvdx0NVeyoJx5qZbxIrJR1dx1kXj6BRhErcESxJnsM3UI/4uAguSZGSbFlWm\nSEWd1cL0/Y2UK3XPfCvxVWdDbA9VFJfVuKm6pSu2Uh+m/VwdaHEdjC736sg7yYSkKF3iutbfsPMg\nrtvyLNZuG25Jw58Z3XmrE9809mYm6Vk9Qd3kows4MFUvzRPh7hXzmn4uW01SZ7DJ1COOB1QFeVQQ\nu6JbVBV7C9aIybjfH2BuZ5AlXrMIiWuKcVmNR5Pt4uJa6kO3X5xQXDFOl1Wby71qxgQXp1fCSLmS\nSc9oGXklfub8OIL8REPHP8GsngAfuuyn8Pzhk86CobcQ4CdnK9Ys7rOKlbjO1q9LCt16+5La9s1U\njU0jmztO+m3SEHaTqSctP10zIeXa789QBaEdjY8Ar1nUiNMXAtD3uVi5cE7dZ/191X7ZLu/FrJ6g\n2gwH6rBUV6eb2C+OABTjfEOTeCTjcq+aiZG39UxuJSqTTWWcMYF6DWJWT4CHVi/F+psW4aWjp50F\nBQEYXv9JY+tcgS2kU4Rpr902jEumd6EgNeEQSaHie/3b/W87jrCRGd15VMa56V4Wd187z3nbpGHl\nOq3ZVtE2Ds0myqq4ZJp+Hd+OxkeA1yxqxF0d9PcVsffNk3VRLwxgx74Sll8529kvIHApTOYa9icm\nlGbrSMUZR/RemQrk2WzmIgOWaDKJbnr4/zQRfm3Tit40ThGVJE8MKl+WCfFdmLohyojGUdFgi2jP\n81OjFRSCPB7S9FBJqo3lc1Qz940zI5+jRNn01109O1ZflaTPsk5rTjsRMK1oLluwRjuTYL1mEZJk\ndaBqG6pb/ZlWynFUVtmmqVvXCXvsyoVzMkl2crlXpvvmOrUwV52gd6+Yh1d+51OY0Z1c04jeq0KQ\nx9c/uxTTu8yvgC1QKvp9x1mZyt+F6yRFQJ3dfuCx/Rh4fL8yV0JEYM1ftwtX3/cU5ocJl/cPHnAe\nY5SoYEhaduWlo6djJZ81k7in0ppbmQgYB5NfMg2NpRm8ZmHBFIER18wDNBYLjBt9Ia9g5HINUcqV\ncTy85yh6CwGmB7nYzeTjjENHMaVeDY++cAzLr5zt5MAO8gQwGjoH3rasWJdENz000dgicibYHqEi\nVvsDqxbEioCTX3xXrbHBJGaZrOVWrGKszXa4S4O4/oG0E/eaPV5WZTp0cwoBmZVEd4W4gyt3xmH5\n8uW8d+/exPur1L8gRwChoXKoeMl1LRCzrHWv4qp1u4yr9XZ12hocKqXWq8FF8BBV7eCqstZAsjBT\noJqtvfvQCaeubNH2ozpm9QQY+mp9s6poK16ZIAd0SLO/VJG7Bk4VdLkZabxj7ZhTiGgfMy+3befN\nUCHKDnQTjTH4stmhU1RZmwkjbq2btOjvK6ZSKypP5LRaZ676jAA0mB100TQuPtrdh07guXXX46HV\nS41O93JlHLsPnbCGP1I41qh5CKg6ux9avbTOxPfQ6qWY4I6r3N80zTrI20XcMPs4dMqcosKboULi\n2JrFtq2saWNCdPcyaRftCrf7zJLLmjZ7fGBOD378z2ecttWZN3TX76JYl0bKdYlQutW/OI9sort/\n8EDD9TOgrCkl92iPjt+koanCWDsl8c5EJ/cjMZFlwc9OmVNUeGEREsfWHM2laPcX2d9XtJp72hVu\nZyqF4IqroBCostd1NYBcu9RFS8nrhEX0Pse5/so4Y+OT6nIOuqZDBGDr7UsaJpe9b57sCN+ECVvy\nWaeSVZShoBPmFBVeWIQom79rfBZpqoRpOcpMNv12qrHt0Gjk7HVTxd84iWWyxmJaVEQLTMZdO4uo\npuhzseIDs5TlU0T/jOgz064s3zhEc5KmChdCX5ckeGERYioYmJVK2Ew/7Chx+ni3kjTqY8VBJEbq\nQhDzRJhgrhVhjJNvIMxRJhPP3jdPGivSuqB6Lk6eOY/rrp6NPa+fquY3SE2MVLTL7BgHF61LtZgC\n2mum6WRTUZb4aKg2knbkQyd23RocKrW8v4UpyokAvBFG4NiiyKLkifCzM6dnKvwKQQ6zZ0xr+rlY\nuvGZzMugNIv8XahQRihqQqPbmX8w1XGNhvKaRRtJ21HWibbO/r5iQ3axDV3dJVfkKqJRensmo7Pi\naj3jzJmv2McmWDsmuQ9HdCEgLxRmFgK8ey6bTntpYrPxKyMUFWbDpDWdgM5cYHUqPnS2jdhqLMkt\nWXWtTjsZMf64ndhsq/3rrp5dXWEaEO1Jo7x3dqx2H+PamIu9hcwDBWw1l1QVV6MVWUfKFWVmdY6q\nOSOd4Fh2sfEniVCMw1Rra9puvLBoI6aY6qn+IMvjj4tOWOSJcM+KeXjkP/5ra8xrsbeAGd2NinNl\ngrFm23Atr8EV8b20wikrymWbKFfGsfHJann2NduGnZILJxh4JIyQukdRplwnotLKhhDHcS1bEUcw\nJxHiWeZLXIh4YdFGTDWWpvqDbOu9Qag6313JE+Frn12CB/oXY3CoZMxmFhO7yWZfGilj7bZhBJo3\noCfIKb+XNEKBbRCA25YVa+fXcWq0ElsYyx0C5XMUewtaIc2o3o9muXvFPBxxqGosUC2mgjxVoxQl\nkkYiZZkvcSHifRZtRudnmOoPsm2cDLeEOME4cy1SzCYwxcSuazsqj6Ey0ZhrEeQJv3frh2N9L2nC\nmMwaT7NkikxlnPHEvuP44e98qvaZLuACALq78mA09oCIg64is46sIxSzzpe40PDCokOZyg/y4FBJ\nmwQnEzdaR2hWpgm7txDUJg7XDOGZhQA93V1Ok09vTxDbB6PDFLVVGilj/rpdiY8d5AiXTO/CyGhF\nqzGMVibqMtNXLpyjTeQbKVfw0OqldU70M+fH6hzOBKDLUMMqiSNat5hKwwl9seZLJMWHznYoWRYr\ny5JWhMqaEhAfWr0UgLkkRxRbCKdMnJBUQlUQqbYX1YZt2k8SCFWTj8jBsAmd3kKADTcvMnZ0zBPh\n8OZP132miiSylZ0BELvScpb4aCgfOjvlmaqJPxt2HsxUUOSJlCtCMUEC6paiJhiolRi33d/TMQTF\n3SuqFXB1q1dxLpfS5HFgTJZ17+8romBpHDVSrljvmUqg6TLHbX6UZpJP06YTw807Fa9ZeFKlGdOJ\nK72R1bpYGZvKxrvgorm5Hl8uKT6jO48gn8PpsrqnyOBQKXYuigsE4Bevno0fHDnlVNZEl5sCuCcE\n3j94oK57pIlWl/L3qPElyj0XLFGzzpnzkwlozTigXaLNXHuDywv5M+fHMVLW+w76+4roUYT5NgsD\neO7wSef6VzpBEeTJyY4/OFTCjn0l56z4qRKs4anizVCeVJmVogPYlcp4NXdiw86DTffqNk1gwr5t\nyhC3IfqLb3zyIG788GXYfehEooKDrSJObTFbuHSUqRCs4ZnEaxaeVFl/0yJrdnVWjJQrTQkKQD+B\nRZMMReJc0uZOp0YreHjP0VrSZSeSJ8Kp0Uqtj7etikAcTSHIuWkrns7BCwtPqvT3FbH19iUdUVIi\nLqawSVOnPRezVCcR5Agzuu1jVvXvNlURmBlHcE7NJnkXNV5YeFKnv6/o1Ia00zA5t3Wr5pHRCjbf\nujj1FqFxj+Z6/lk9AbryhDPnq4IvF/NEJr9OnFtQGecpU43AU8ULC09myOVMdIT9pWKR1aJ0zbZh\n9G16RrlyNhV97O8r4mufXZKK+U303J4eQ8gWgjzuunauVTD3BDm8d3aszlSXJMrZJDjTOI6nM/HC\nwpMpspYRnUyDPOHrn12KB8NEOhcKQR4Prl6qPF4anBqtYODx/Q0Cw1T0EZg0v8WpdxVFHC+Oo1jU\nrXqgf7FVMI9WJrS5FPKdtGkbtmrJrngH99TCR0N5MiXaZ4GougKV8w1MFWBn9ZhLccj5Cb2FAJ9Z\nchkefeFYU1nRwkQin8clSTKa4BXNDp5/aQHPHz5Z59AWXffkrOa1DrWgVDkh4vxJc12OhFnsquoB\n8nl1fp2BVQuc61j5shpTDy8sPJkRnXRGypWaZiBPciZzhClsU5d9+4imvpGMqS6Tbkxxs31V27uU\nl9DVBZNbwpqyzXXhy0T64o2yc1oWjKWRci1M2Famw9Toyib0PZ2PFxaezDCVWZcnCt3kKBcFjIOt\nA57IHTCVpmjWRKITCi4CR1fgzrUu2PqbFmHg8f11yXhBnrD6o3O1/cHPnB+rKyqYtAzG+psWKcfe\nzj7wnnTwwsKTGa5l1lcunNNQIqIQ5LHh5kUN+6omYaDePLRy4Rzs2FfSag493V21iUtVE8k1Y1lH\nVKNyrYUUNdlND3INJjvdPVCZzORV/ozuLiy/cjaWXzkba7cPN2gYKtNbnOuVx3PbsmIt2dBrERcO\nXlh4MsOlzLqqRIRo/qMy4UQn4YHH9wOM2oRfGiljx74SbltW1JbbFsJKHF+uUBsnY1mHq0Zlujad\nyS6OIDorRT2NlCu474kD2HzrYm0rwmZak8rj2bGv1PHVkT3x8cLCkxku/QJUE6to/hNFta2q7lG5\nMo7dh05oS5nLwiqLqqNJGle5Cpg0tkuzV0oSweiZmmQaOktENxDRq0T0GhGtU/x9GhFtC//+AhHN\nl/52X/j5q0S0KstxerLB1DZWEGdijbPyfWukbA13zYokoaWu9yGN7dK8L1O9o6PHncw0CyLKA/hD\nAJ8AcBzAi0S0k5l/KG32BQCnmPnniOhOAL8PYDURfQjAnQAWAbgcwN8R0c8zc3pF/z0twbZyj7PK\ntTmuo9u2qydIkg5srvchje3SvC9TuaOjJx5ZahbXAHiNmV9n5vMAvg3glsg2twD4y/DnxwH8ChFR\n+Pm3mfkcM78B4LXweJ4LjDirXNW2QZ4QRLLIoslyz627Hm9suRHPrbu+JaYRF40qiut9SGu7tO5L\nu7Q3T+vJ0mdRBHBM+v04gGt12zDzGBGdBnBp+PmeyL4NTzMRfRHAFwFg3rx5qQ3c0zrirHJ127ru\n30qS5GQA9utIe7tmmaodHT3xyaxTHhHdAWAVM/+H8Pf/HcA1zPyfpG0OhtscD38/jKoGsQnAPzDz\nw+HnfwbgKWbeoTuf75Tn8Xg88emETnnHAcyVfr8CwFu6bYioC8BMACcd9/V4PB5Pi8hSWLwI4INE\ndBURdaPqsN4Z2WYngM+HP98O4Fmuqjo7AdwZRktdBeCDAH6Q4Vg9Ho/HYyAzn0Xog/gNAE8DyAP4\nFjMfJKJNAPYy804Afwbgr4noNVQ1ijvDfQ8S0XYAPwQwBuDXfSSUx+PxtI/MfBatxvssPB6PJz6d\n4LPweDwezwWCFxYej8fjseKFhcfj8XiseGHh8Xg8HiteWHg8Ho/HihcWHo/H47HihYXH4/F4rHhh\n4fF4PB4rXlh4PB6Px4oXFh6Px+Ox4oWFx+PxeKx4YeHxeDweK15YeDwej8eKFxYej8fjsXLBlCgn\nohMA3ky4+/sB/EuKw8kCP8bm6fTxAX6MadHpY+yk8V3JzHNsG10wwqIZiGivSz33duLH2DydPj7A\njzEtOn2MnT4+Fd4M5fF4PB4rXlh4PB6Px4oXFlW+2e4BOODH2DydPj7AjzEtOn2MnT6+BrzPwuPx\neDxWvGbh8Xg8HiteWHg8Ho/HykUtLIhoKxEdIqKXieg7RNQr/e0+InqNiF4lolVtGt8dRHSQiCaI\naHnkb20fnzSWG8JxvEZE69o5FgERfYuI/pmI/lH6bDYRfY+Ifhz+f1abxziXiHYT0Svh9/ybnTRO\nIppORD8gov3h+DaGn19FRC+E49tGRN3tGF9krHkiGiKiv+3EMRLRESI6QETDRLQ3/KwjvmdXLmph\nAeB7AP4VM38YwI8A3AcARPQhAHcCWATgBgD/LxHl2zC+fwRwK4D/KX/YQeNDeN4/BPApAB8CcFc4\nvnbzF6jeG5l1AP6emT8I4O/D39vJGIB7mfkXAKwA8OvhveuUcZ4DcD0zLwGwFMANRLQCwO8DeDAc\n3ykAX2jT+GR+E8Ar0u+dOMaVzLxUyq/olO/ZiYtaWDDzM8w8Fv66B8AV4c+3APg2M59j5jcAvAbg\nmjaM7xVmflXxp44YX8g1AF5j5teZ+TyAb4fjayvM/D8BnIx8fAuAvwx//ksA/S0dVARmfpuZXwp/\nfhfVya6IDhknV3kv/DUI/zGA6wE8Hn7e9vtIRFcAuBHAn4a/EzpsjBo64nt25aIWFhH+PYDvhj8X\nARyT/nY8/KxT6KTxddJYbPwMM78NVCdqAD/d5vHUIKL5APoAvIAOGmdo3hkG8M+oauKHAYxIi6xO\n+L4fAvBfAEyEv1+KzhsjA3iGiPYR0RfDzzrme3ahq90DyBoi+jsAP6v401eY+b+H23wFVZPAI2I3\nxfaZxBi7jE+1m+KzdsVAd9JYpiREdAmAHQDWMPNPqgvjzoCZxwEsDf153wHwC6rNWjuqSYjoMwD+\nmZn3EdHHxMeKTdv9TF7HzG8R0U8D+B4RHWrzeGJzwQsLZv646e9E9HkAnwHwKzyZdHIcwFxpsysA\nvNWO8Wlo2fim2Fhs/BMRXcbMbxPRZaiultsKEQWoCopHmPmJ8OOOGyczj9D/3969hMZVR3Ec//6g\n2GhtCYqIT1K1SkGCLsxGEYUQNItIsYqoUKzooqAUqfgCKRR8oHThwhcqBVEXXYihRQK2cVOUVIxN\nU4r4rtCi+EBMLUXa4+L8gzchyU0a6B06vw9cMnNn8p8zuTNz7v3fyTnSp+S5lU5JS8qee9Pb+0Zg\nQFI/0AGsII80WilGIuJw+fmrpA/J6duW285zaetpKEm3AU8AAxHxT+WmQeAeSUslrQRWASNNxDiL\nVopvL7CqfPvkLPLE+2BDsdQZBNaVy+uA2Y7cTosyt/42cDAitlZuaok4JV0w+Q1BSWcDveR5lWFg\nbdPxAUTEUxFxaUR0ka+93RFxHy0Uo6RlkpZPXgb6yC+vtMR2nreIaNuFPDH8M/BVWV6v3PYMOT/7\nNXB7Q/GtIffcjwO/AEOtFF8lln7y22TfkdNnrbBtPwCOAP+Wv+GD5Fz2LuCb8vO8hmO8iZweGau8\nBvtbJU6gGxgt8Y0Dz5b1V5A7J98C24GlTW/vEtctwI5Wi7HEsq8sBybfI62ynee7uNyHmZnVautp\nKDMzmx8nCzMzq+VkYWZmtZwszMyslpOFmZnVcrKwtiGpU9KGpuOoI2mjpHOajsOsysnC2kkn0Hiy\nUJrrvbcRWFCykHTGV2OwZjlZWDt5Abiy9BR4SdLjkvYq+5lM9mroUvY4eUvSuKT3JPVK2lP6DvSU\n+22W9K6k3WX9Q5MPMse4ByW9CnwJXCbpNUlfTOsV8ShwMTAsabism6iMvVbStnJ5m6St5X4vlv8U\nfqc89qikxqv/2pnDeyPWTp4k+5dcJ6mPLAfRQxaeG5R0M3AIuAq4C3iYLGdyL/nf1gPA0/xfSrqb\nrJW0DBiVtBO4n+jRvQAAAbdJREFUliy/MtO41wAPRMQGyAKWEfFH6QmyS1J3RLwi6TGy98Fv83hO\nVwO9EXFC0nNkuYv1pUzHiKRPIuLoIv5mZoCThbWvvrKMluvnkh/yh4AfImI/gKQDZIOakLQf6KqM\n8VFEHAOOlb37HjKpzDbuTxHxeeX37y7lqpcAF5HNo8YW+Dy2R1aGnXxOA5I2lesdwOVMbQpkdkqc\nLKxdCXg+It6YsjL7ShyvrDpZuX6Sqe+Z6bVyombco5XrK4FNwA0R8WeZWuqYJdbq40y/T/WoQcCd\nMXPDLLNF8TkLayd/A8vL5SFgfeklgaRLSq+BhbhD2af6fLKI3d4FjLuC/KD/S9KFZFvameKELGW9\nupwUXzNHPEPAI6WaLZKuX+DzMZuVjyysbUTE7+VE9TjZFfF94LPy2ToB3A+cmGOI6UaAneRUz5bI\nngWHJa2uGzci9kkaJauQfg/sqdz8JvCxpCMRcSt5rmUHWSF5nJzamskWspfDWEkYP5K9WswWzVVn\nzU6BpM3ARES83HQsZqeDp6HMzKyWjyzMzKyWjyzMzKyWk4WZmdVysjAzs1pOFmZmVsvJwszMav0H\nuwGqeLgOTX4AAAAASUVORK5CYII=\n", | |
| "text/plain": [ | |
| "<matplotlib.figure.Figure at 0x1c13d44390>" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "output_type": "display_data" | |
| } | |
| ], | |
| "source": [ | |
| "x = np.array(df['temperature'])\n", | |
| "y = np.array(df['nitrate'])\n", | |
| "plt.scatter(x, y)\n", | |
| "plt.title('temperature vs nitrate')\n", | |
| "plt.xlabel('temperature')\n", | |
| "plt.ylabel('nitrate')\n", | |
| "plt.show()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "collapsed": true | |
| }, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "\n", | |
| "There are clearly higher nitrate levels at more extreme temperatures. Let's make a binary variable out of this, where 1 indicates an extreme temperature and 0 indicates a regular temperature." | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 9, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<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>Unnamed: 0</th>\n", | |
| " <th>feed</th>\n", | |
| " <th>temperature</th>\n", | |
| " <th>nitrate</th>\n", | |
| " <th>extreme_temp</th>\n", | |
| " </tr>\n", | |
| " </thead>\n", | |
| " <tbody>\n", | |
| " <tr>\n", | |
| " <th>0</th>\n", | |
| " <td>0</td>\n", | |
| " <td>5.60</td>\n", | |
| " <td>25.62</td>\n", | |
| " <td>0.102</td>\n", | |
| " <td>0</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>1</th>\n", | |
| " <td>1</td>\n", | |
| " <td>9.87</td>\n", | |
| " <td>29.39</td>\n", | |
| " <td>0.042</td>\n", | |
| " <td>0</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>2</th>\n", | |
| " <td>2</td>\n", | |
| " <td>5.29</td>\n", | |
| " <td>14.92</td>\n", | |
| " <td>0.028</td>\n", | |
| " <td>0</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>3</th>\n", | |
| " <td>3</td>\n", | |
| " <td>9.49</td>\n", | |
| " <td>31.16</td>\n", | |
| " <td>0.031</td>\n", | |
| " <td>0</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>4</th>\n", | |
| " <td>4</td>\n", | |
| " <td>6.01</td>\n", | |
| " <td>31.61</td>\n", | |
| " <td>0.075</td>\n", | |
| " <td>0</td>\n", | |
| " </tr>\n", | |
| " </tbody>\n", | |
| "</table>\n", | |
| "</div>" | |
| ], | |
| "text/plain": [ | |
| " Unnamed: 0 feed temperature nitrate extreme_temp\n", | |
| "0 0 5.60 25.62 0.102 0\n", | |
| "1 1 9.87 29.39 0.042 0\n", | |
| "2 2 5.29 14.92 0.028 0\n", | |
| "3 3 9.49 31.16 0.031 0\n", | |
| "4 4 6.01 31.61 0.075 0" | |
| ] | |
| }, | |
| "execution_count": 9, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df['extreme_temp'] = [1 if x <= 0 or x >= 35 else 0 for x in df['temperature']]\n", | |
| "df.head()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 6: Rerun Regression with New Variable" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Using statsmodels" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 10, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<table class=\"simpletable\">\n", | |
| "<caption>OLS Regression Results</caption>\n", | |
| "<tr>\n", | |
| " <th>Dep. Variable:</th> <td>feed</td> <th> R-squared: </th> <td> 0.666</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Model:</th> <td>OLS</td> <th> Adj. R-squared: </th> <td> 0.665</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Method:</th> <td>Least Squares</td> <th> F-statistic: </th> <td> 928.3</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Date:</th> <td>Sat, 21 Mar 2020</td> <th> Prob (F-statistic):</th> <td> 0.00</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Time:</th> <td>17:37:41</td> <th> Log-Likelihood: </th> <td> -3723.2</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>No. Observations:</th> <td> 1400</td> <th> AIC: </th> <td> 7452.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Residuals:</th> <td> 1397</td> <th> BIC: </th> <td> 7468.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Model:</th> <td> 3</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Covariance Type:</th> <td>nonrobust</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <td></td> <th>coef</th> <th>std err</th> <th>t</th> <th>P>|t|</th> <th>[0.025</th> <th>0.975]</th> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>nitrate</th> <td> 70.6729</td> <td> 1.913</td> <td> 36.940</td> <td> 0.000</td> <td> 66.920</td> <td> 74.426</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>temperature</th> <td> 0.0541</td> <td> 0.004</td> <td> 12.366</td> <td> 0.000</td> <td> 0.046</td> <td> 0.063</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>extreme_temp</th> <td> -6.2473</td> <td> 0.235</td> <td> -26.633</td> <td> 0.000</td> <td> -6.707</td> <td> -5.787</td>\n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <th>Omnibus:</th> <td>20.706</td> <th> Durbin-Watson: </th> <td> 1.699</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Prob(Omnibus):</th> <td> 0.000</td> <th> Jarque-Bera (JB): </th> <td> 21.135</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Skew:</th> <td>-0.288</td> <th> Prob(JB): </th> <td>2.57e-05</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Kurtosis:</th> <td> 2.825</td> <th> Cond. No. </th> <td> 551.</td>\n", | |
| "</tr>\n", | |
| "</table>" | |
| ], | |
| "text/plain": [ | |
| "<class 'statsmodels.iolib.summary.Summary'>\n", | |
| "\"\"\"\n", | |
| " OLS Regression Results \n", | |
| "==============================================================================\n", | |
| "Dep. Variable: feed R-squared: 0.666\n", | |
| "Model: OLS Adj. R-squared: 0.665\n", | |
| "Method: Least Squares F-statistic: 928.3\n", | |
| "Date: Sat, 21 Mar 2020 Prob (F-statistic): 0.00\n", | |
| "Time: 17:37:41 Log-Likelihood: -3723.2\n", | |
| "No. Observations: 1400 AIC: 7452.\n", | |
| "Df Residuals: 1397 BIC: 7468.\n", | |
| "Df Model: 3 \n", | |
| "Covariance Type: nonrobust \n", | |
| "================================================================================\n", | |
| " coef std err t P>|t| [0.025 0.975]\n", | |
| "--------------------------------------------------------------------------------\n", | |
| "nitrate 70.6729 1.913 36.940 0.000 66.920 74.426\n", | |
| "temperature 0.0541 0.004 12.366 0.000 0.046 0.063\n", | |
| "extreme_temp -6.2473 0.235 -26.633 0.000 -6.707 -5.787\n", | |
| "==============================================================================\n", | |
| "Omnibus: 20.706 Durbin-Watson: 1.699\n", | |
| "Prob(Omnibus): 0.000 Jarque-Bera (JB): 21.135\n", | |
| "Skew: -0.288 Prob(JB): 2.57e-05\n", | |
| "Kurtosis: 2.825 Cond. No. 551.\n", | |
| "==============================================================================\n", | |
| "\n", | |
| "Warnings:\n", | |
| "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n", | |
| "\"\"\"" | |
| ] | |
| }, | |
| "execution_count": 10, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "X = df[['nitrate', 'temperature', 'extreme_temp']]\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "model.summary()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "\n", | |
| "We can interpret the results as follows:\n", | |
| "- The model is significant as the F-statistic is high and the p-value for the statistic is also less than 0.05\n", | |
| "- All three features are statistically significant as their p-values are less than 0.05\n", | |
| "- Nitrate and Temperature are positively correlated with feed\n", | |
| "- Extreme_temp is negatively correlated with feed. Given that extreme_temp is a binary variable, this means that more data points were at a regular temperature (0) than at an extreme temperature (1).\n", | |
| "- **The adjusted R^2 value is significantly higher than before!** That means that we have made a model that explains variability in data much better" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 7: Plot the Residuals\n", | |
| "\n", | |
| "Residuals plot the difference between fitted values (the predictions) and the actual values. The center dotted line represents (in this case) the best-fit line. The data points around the line indicate the actual values and how far away from the best-fit line they are." | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### We first need to get the predicted values" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 11, | |
| "metadata": { | |
| "collapsed": true | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "X = df[['nitrate', 'temperature', 'extreme_temp']]\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "predictions = model.predict(X)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 12, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEKCAYAAAASByJ7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/h96pKn4SvjUKwAAAABJRU5ErkJggg==\n", | |
| "text/plain": [ | |
| "<matplotlib.figure.Figure at 0x1c13cce518>" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "output_type": "display_data" | |
| } | |
| ], | |
| "source": [ | |
| "sns.residplot(predictions, y)\n", | |
| "plt.show()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "Looking at the above, there is non-linearity in the residuals. In specific, there appears to be a downward curve to the residuals." | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 13, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEKCAYAAAASByJ7AAAAAXNSR0IArs4c6QAAAAlwSFlzAAAL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\nVlLvaAAAAABJRU5ErkJggg==\n", | |
| "text/plain": [ | |
| "<IPython.core.display.Image object>" | |
| ] | |
| }, | |
| "execution_count": 13, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "from IPython.display import Image\n", | |
| "Image('residuals_case.png')" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "This indicates that we are likely not using an important non-linear independent variable.\n", | |
| "\n", | |
| "Because of the shape of the curve, it appears we need to add a log(X) feature. Which feature should we transform?" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 14, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<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>Unnamed: 0</th>\n", | |
| " <th>feed</th>\n", | |
| " <th>temperature</th>\n", | |
| " <th>nitrate</th>\n", | |
| " <th>extreme_temp</th>\n", | |
| " </tr>\n", | |
| " </thead>\n", | |
| " <tbody>\n", | |
| " <tr>\n", | |
| " <th>count</th>\n", | |
| " <td>1400.0000</td>\n", | |
| " <td>1400.000000</td>\n", | |
| " <td>1400.000000</td>\n", | |
| " <td>1400.000000</td>\n", | |
| " <td>1400.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>mean</th>\n", | |
| " <td>699.5000</td>\n", | |
| " <td>5.274100</td>\n", | |
| " <td>17.424721</td>\n", | |
| " <td>0.080006</td>\n", | |
| " <td>0.428571</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>std</th>\n", | |
| " <td>404.2895</td>\n", | |
| " <td>2.823031</td>\n", | |
| " <td>19.997534</td>\n", | |
| " <td>0.041928</td>\n", | |
| " <td>0.495048</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>min</th>\n", | |
| " <td>0.0000</td>\n", | |
| " <td>0.000000</td>\n", | |
| " <td>-19.850000</td>\n", | |
| " <td>0.003000</td>\n", | |
| " <td>0.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>25%</th>\n", | |
| " <td>349.7500</td>\n", | |
| " <td>2.930000</td>\n", | |
| " <td>2.737500</td>\n", | |
| " <td>0.047000</td>\n", | |
| " <td>0.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>50%</th>\n", | |
| " <td>699.5000</td>\n", | |
| " <td>5.540000</td>\n", | |
| " <td>16.965000</td>\n", | |
| " <td>0.077000</td>\n", | |
| " <td>0.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>75%</th>\n", | |
| " <td>1049.2500</td>\n", | |
| " <td>7.682500</td>\n", | |
| " <td>32.292500</td>\n", | |
| " <td>0.107000</td>\n", | |
| " <td>1.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>max</th>\n", | |
| " <td>1399.0000</td>\n", | |
| " <td>9.970000</td>\n", | |
| " <td>54.790000</td>\n", | |
| " <td>0.195000</td>\n", | |
| " <td>1.000000</td>\n", | |
| " </tr>\n", | |
| " </tbody>\n", | |
| "</table>\n", | |
| "</div>" | |
| ], | |
| "text/plain": [ | |
| " Unnamed: 0 feed temperature nitrate extreme_temp\n", | |
| "count 1400.0000 1400.000000 1400.000000 1400.000000 1400.000000\n", | |
| "mean 699.5000 5.274100 17.424721 0.080006 0.428571\n", | |
| "std 404.2895 2.823031 19.997534 0.041928 0.495048\n", | |
| "min 0.0000 0.000000 -19.850000 0.003000 0.000000\n", | |
| "25% 349.7500 2.930000 2.737500 0.047000 0.000000\n", | |
| "50% 699.5000 5.540000 16.965000 0.077000 0.000000\n", | |
| "75% 1049.2500 7.682500 32.292500 0.107000 1.000000\n", | |
| "max 1399.0000 9.970000 54.790000 0.195000 1.000000" | |
| ] | |
| }, | |
| "execution_count": 14, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df.describe()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "Looking at the above, the only feature we should try to apply a log transformation on is the nitrate feature.\n", | |
| "- This is because temperature has negative values (see minimum value), which we cannot take the log of\n", | |
| "- The minimum value for nitrate is greater than zero, which means that all nitrate values are positive and we can take the log of it" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 8: Rerun Regression with New Log Variable" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Add log_nitrate feature" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 15, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<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>Unnamed: 0</th>\n", | |
| " <th>feed</th>\n", | |
| " <th>temperature</th>\n", | |
| " <th>nitrate</th>\n", | |
| " <th>extreme_temp</th>\n", | |
| " <th>log_nitrate</th>\n", | |
| " </tr>\n", | |
| " </thead>\n", | |
| " <tbody>\n", | |
| " <tr>\n", | |
| " <th>0</th>\n", | |
| " <td>0</td>\n", | |
| " <td>5.60</td>\n", | |
| " <td>25.62</td>\n", | |
| " <td>0.102</td>\n", | |
| " <td>0</td>\n", | |
| " <td>-2.282782</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>1</th>\n", | |
| " <td>1</td>\n", | |
| " <td>9.87</td>\n", | |
| " <td>29.39</td>\n", | |
| " <td>0.042</td>\n", | |
| " <td>0</td>\n", | |
| " <td>-3.170086</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>2</th>\n", | |
| " <td>2</td>\n", | |
| " <td>5.29</td>\n", | |
| " <td>14.92</td>\n", | |
| " <td>0.028</td>\n", | |
| " <td>0</td>\n", | |
| " <td>-3.575551</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>3</th>\n", | |
| " <td>3</td>\n", | |
| " <td>9.49</td>\n", | |
| " <td>31.16</td>\n", | |
| " <td>0.031</td>\n", | |
| " <td>0</td>\n", | |
| " <td>-3.473768</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>4</th>\n", | |
| " <td>4</td>\n", | |
| " <td>6.01</td>\n", | |
| " <td>31.61</td>\n", | |
| " <td>0.075</td>\n", | |
| " <td>0</td>\n", | |
| " <td>-2.590267</td>\n", | |
| " </tr>\n", | |
| " </tbody>\n", | |
| "</table>\n", | |
| "</div>" | |
| ], | |
| "text/plain": [ | |
| " Unnamed: 0 feed temperature nitrate extreme_temp log_nitrate\n", | |
| "0 0 5.60 25.62 0.102 0 -2.282782\n", | |
| "1 1 9.87 29.39 0.042 0 -3.170086\n", | |
| "2 2 5.29 14.92 0.028 0 -3.575551\n", | |
| "3 3 9.49 31.16 0.031 0 -3.473768\n", | |
| "4 4 6.01 31.61 0.075 0 -2.590267" | |
| ] | |
| }, | |
| "execution_count": 15, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df['log_nitrate'] = np.log(df['nitrate'])\n", | |
| "df.head()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Re-run model with log_nitrate feature" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 16, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<table class=\"simpletable\">\n", | |
| "<caption>OLS Regression Results</caption>\n", | |
| "<tr>\n", | |
| " <th>Dep. Variable:</th> <td>feed</td> <th> R-squared: </th> <td> 0.936</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Model:</th> <td>OLS</td> <th> Adj. R-squared: </th> <td> 0.936</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Method:</th> <td>Least Squares</td> <th> F-statistic: </th> <td> 5112.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Date:</th> <td>Sat, 21 Mar 2020</td> <th> Prob (F-statistic):</th> <td> 0.00</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Time:</th> <td>17:37:43</td> <th> Log-Likelihood: </th> <td> -2565.5</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>No. Observations:</th> <td> 1400</td> <th> AIC: </th> <td> 5139.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Residuals:</th> <td> 1396</td> <th> BIC: </th> <td> 5160.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Model:</th> <td> 4</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Covariance Type:</th> <td>nonrobust</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <td></td> <th>coef</th> <th>std err</th> <th>t</th> <th>P>|t|</th> <th>[0.025</th> <th>0.975]</th> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>nitrate</th> <td> 30.1443</td> <td> 0.990</td> <td> 30.464</td> <td> 0.000</td> <td> 28.203</td> <td> 32.085</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>temperature</th> <td> 0.0027</td> <td> 0.002</td> <td> 1.349</td> <td> 0.178</td> <td> -0.001</td> <td> 0.007</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>extreme_temp</th> <td> -5.0629</td> <td> 0.104</td> <td> -48.781</td> <td> 0.000</td> <td> -5.267</td> <td> -4.859</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>log_nitrate</th> <td> -1.8301</td> <td> 0.024</td> <td> -76.817</td> <td> 0.000</td> <td> -1.877</td> <td> -1.783</td>\n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <th>Omnibus:</th> <td>175.599</td> <th> Durbin-Watson: </th> <td> 1.977</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Prob(Omnibus):</th> <td> 0.000</td> <th> Jarque-Bera (JB): </th> <td> 46.585</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Skew:</th> <td>-0.048</td> <th> Prob(JB): </th> <td>7.66e-11</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Kurtosis:</th> <td> 2.111</td> <th> Cond. No. </th> <td> 652.</td>\n", | |
| "</tr>\n", | |
| "</table>" | |
| ], | |
| "text/plain": [ | |
| "<class 'statsmodels.iolib.summary.Summary'>\n", | |
| "\"\"\"\n", | |
| " OLS Regression Results \n", | |
| "==============================================================================\n", | |
| "Dep. Variable: feed R-squared: 0.936\n", | |
| "Model: OLS Adj. R-squared: 0.936\n", | |
| "Method: Least Squares F-statistic: 5112.\n", | |
| "Date: Sat, 21 Mar 2020 Prob (F-statistic): 0.00\n", | |
| "Time: 17:37:43 Log-Likelihood: -2565.5\n", | |
| "No. Observations: 1400 AIC: 5139.\n", | |
| "Df Residuals: 1396 BIC: 5160.\n", | |
| "Df Model: 4 \n", | |
| "Covariance Type: nonrobust \n", | |
| "================================================================================\n", | |
| " coef std err t P>|t| [0.025 0.975]\n", | |
| "--------------------------------------------------------------------------------\n", | |
| "nitrate 30.1443 0.990 30.464 0.000 28.203 32.085\n", | |
| "temperature 0.0027 0.002 1.349 0.178 -0.001 0.007\n", | |
| "extreme_temp -5.0629 0.104 -48.781 0.000 -5.267 -4.859\n", | |
| "log_nitrate -1.8301 0.024 -76.817 0.000 -1.877 -1.783\n", | |
| "==============================================================================\n", | |
| "Omnibus: 175.599 Durbin-Watson: 1.977\n", | |
| "Prob(Omnibus): 0.000 Jarque-Bera (JB): 46.585\n", | |
| "Skew: -0.048 Prob(JB): 7.66e-11\n", | |
| "Kurtosis: 2.111 Cond. No. 652.\n", | |
| "==============================================================================\n", | |
| "\n", | |
| "Warnings:\n", | |
| "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n", | |
| "\"\"\"" | |
| ] | |
| }, | |
| "execution_count": 16, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "X = df[['nitrate', 'temperature', 'extreme_temp', 'log_nitrate']]\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "model.summary()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "\n", | |
| "We can interpret the above as follows:\n", | |
| "- There is a big increase in the Adjusted R^2 value and the model remains statistically significant as the p-value is less than 0.05 for the high F-statistic\n", | |
| "- In fact, the Adj. R^2 value is nearly 1! Which means now nearly all variability of the data is explained by this model\n", | |
| "- Interestingly, the p-value for temperature is now greater than 0.05, which means it is not significant\n", | |
| "\n", | |
| "Let's try removing the temperature variable now that it is no longer significant in this model." | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 9: Rerun Regression with Reduced Feature Set" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 17, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<table class=\"simpletable\">\n", | |
| "<caption>OLS Regression Results</caption>\n", | |
| "<tr>\n", | |
| " <th>Dep. Variable:</th> <td>feed</td> <th> R-squared: </th> <td> 0.936</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Model:</th> <td>OLS</td> <th> Adj. R-squared: </th> <td> 0.936</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Method:</th> <td>Least Squares</td> <th> F-statistic: </th> <td> 6811.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Date:</th> <td>Sat, 21 Mar 2020</td> <th> Prob (F-statistic):</th> <td> 0.00</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Time:</th> <td>17:37:44</td> <th> Log-Likelihood: </th> <td> -2566.4</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>No. Observations:</th> <td> 1400</td> <th> AIC: </th> <td> 5139.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Residuals:</th> <td> 1397</td> <th> BIC: </th> <td> 5155.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Model:</th> <td> 3</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Covariance Type:</th> <td>nonrobust</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <td></td> <th>coef</th> <th>std err</th> <th>t</th> <th>P>|t|</th> <th>[0.025</th> <th>0.975]</th> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>nitrate</th> <td> 30.4056</td> <td> 0.971</td> <td> 31.325</td> <td> 0.000</td> <td> 28.501</td> <td> 32.310</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>extreme_temp</th> <td> -5.0676</td> <td> 0.104</td> <td> -48.839</td> <td> 0.000</td> <td> -5.271</td> <td> -4.864</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>log_nitrate</th> <td> -1.8407</td> <td> 0.022</td> <td> -81.815</td> <td> 0.000</td> <td> -1.885</td> <td> -1.797</td>\n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <th>Omnibus:</th> <td>179.156</td> <th> Durbin-Watson: </th> <td> 1.971</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Prob(Omnibus):</th> <td> 0.000</td> <th> Jarque-Bera (JB): </th> <td> 47.048</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Skew:</th> <td>-0.049</td> <th> Prob(JB): </th> <td>6.08e-11</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Kurtosis:</th> <td> 2.107</td> <th> Cond. No. </th> <td> 67.7</td>\n", | |
| "</tr>\n", | |
| "</table>" | |
| ], | |
| "text/plain": [ | |
| "<class 'statsmodels.iolib.summary.Summary'>\n", | |
| "\"\"\"\n", | |
| " OLS Regression Results \n", | |
| "==============================================================================\n", | |
| "Dep. Variable: feed R-squared: 0.936\n", | |
| "Model: OLS Adj. R-squared: 0.936\n", | |
| "Method: Least Squares F-statistic: 6811.\n", | |
| "Date: Sat, 21 Mar 2020 Prob (F-statistic): 0.00\n", | |
| "Time: 17:37:44 Log-Likelihood: -2566.4\n", | |
| "No. Observations: 1400 AIC: 5139.\n", | |
| "Df Residuals: 1397 BIC: 5155.\n", | |
| "Df Model: 3 \n", | |
| "Covariance Type: nonrobust \n", | |
| "================================================================================\n", | |
| " coef std err t P>|t| [0.025 0.975]\n", | |
| "--------------------------------------------------------------------------------\n", | |
| "nitrate 30.4056 0.971 31.325 0.000 28.501 32.310\n", | |
| "extreme_temp -5.0676 0.104 -48.839 0.000 -5.271 -4.864\n", | |
| "log_nitrate -1.8407 0.022 -81.815 0.000 -1.885 -1.797\n", | |
| "==============================================================================\n", | |
| "Omnibus: 179.156 Durbin-Watson: 1.971\n", | |
| "Prob(Omnibus): 0.000 Jarque-Bera (JB): 47.048\n", | |
| "Skew: -0.049 Prob(JB): 6.08e-11\n", | |
| "Kurtosis: 2.107 Cond. No. 67.7\n", | |
| "==============================================================================\n", | |
| "\n", | |
| "Warnings:\n", | |
| "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n", | |
| "\"\"\"" | |
| ] | |
| }, | |
| "execution_count": 17, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "X = df[['nitrate', 'extreme_temp', 'log_nitrate']]\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "model.summary()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "\n", | |
| "We can interpret the above as follows:\n", | |
| "- The Adj. R^2 value had negligble change, which means that the continuous temperature variable was not contributing to the model\n", | |
| "- This means that it is only important for us to note whether or not a temperature is within an extreme range" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 10: Re-plot the Residuals" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 18, | |
| "metadata": { | |
| "collapsed": true | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "X = df[['nitrate', 'extreme_temp', 'log_nitrate']]\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "predictions = model.predict(X)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 19, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEKCAYAAAASByJ7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nElNCEBERQAlBRERiSggiIgLA/wG9yCGWJU8v9AAAAABJRU5ErkJggg==\n", | |
| "text/plain": [ | |
| "<matplotlib.figure.Figure at 0x1c1cba6f98>" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "output_type": "display_data" | |
| } | |
| ], | |
| "source": [ | |
| "sns.residplot(predictions, y)\n", | |
| "plt.show()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "\n", | |
| "The residuals now have a more random pattern, though they appear in two main clusters. Recall that we are working with a binary variable (extreme_temp) that indicates whether or not the temperature reading of a bioreactor is in an extreme state.\n", | |
| "- The two clusters may indicate that we get significantly different feed output results in each range\n", | |
| "- Let's verify this before we consider making more changes to the model" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 11: Investigate Clusters" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 20, | |
| "metadata": { | |
| "collapsed": true | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "df_extreme = df[df['extreme_temp'] == 1]\n", | |
| "df_regular = df[df['extreme_temp'] == 0]" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 21, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<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>Unnamed: 0</th>\n", | |
| " <th>feed</th>\n", | |
| " <th>temperature</th>\n", | |
| " <th>nitrate</th>\n", | |
| " <th>extreme_temp</th>\n", | |
| " <th>log_nitrate</th>\n", | |
| " </tr>\n", | |
| " </thead>\n", | |
| " <tbody>\n", | |
| " <tr>\n", | |
| " <th>count</th>\n", | |
| " <td>600.000000</td>\n", | |
| " <td>600.000000</td>\n", | |
| " <td>600.000000</td>\n", | |
| " <td>600.000000</td>\n", | |
| " <td>600.0</td>\n", | |
| " <td>600.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>mean</th>\n", | |
| " <td>715.301667</td>\n", | |
| " <td>2.471633</td>\n", | |
| " <td>17.616517</td>\n", | |
| " <td>0.109882</td>\n", | |
| " <td>1.0</td>\n", | |
| " <td>-2.280787</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>std</th>\n", | |
| " <td>405.232698</td>\n", | |
| " <td>1.450937</td>\n", | |
| " <td>28.350360</td>\n", | |
| " <td>0.037981</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>0.405542</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>min</th>\n", | |
| " <td>6.000000</td>\n", | |
| " <td>0.000000</td>\n", | |
| " <td>-19.850000</td>\n", | |
| " <td>0.023000</td>\n", | |
| " <td>1.0</td>\n", | |
| " <td>-3.772261</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>25%</th>\n", | |
| " <td>361.750000</td>\n", | |
| " <td>1.207500</td>\n", | |
| " <td>-10.385000</td>\n", | |
| " <td>0.082750</td>\n", | |
| " <td>1.0</td>\n", | |
| " <td>-2.491945</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>50%</th>\n", | |
| " <td>734.500000</td>\n", | |
| " <td>2.440000</td>\n", | |
| " <td>17.520000</td>\n", | |
| " <td>0.111000</td>\n", | |
| " <td>1.0</td>\n", | |
| " <td>-2.198225</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>75%</th>\n", | |
| " <td>1070.250000</td>\n", | |
| " <td>3.752500</td>\n", | |
| " <td>45.835000</td>\n", | |
| " <td>0.138250</td>\n", | |
| " <td>1.0</td>\n", | |
| " <td>-1.978697</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>max</th>\n", | |
| " <td>1399.000000</td>\n", | |
| " <td>4.990000</td>\n", | |
| " <td>54.790000</td>\n", | |
| " <td>0.195000</td>\n", | |
| " <td>1.0</td>\n", | |
| " <td>-1.634756</td>\n", | |
| " </tr>\n", | |
| " </tbody>\n", | |
| "</table>\n", | |
| "</div>" | |
| ], | |
| "text/plain": [ | |
| " Unnamed: 0 feed temperature nitrate extreme_temp \\\n", | |
| "count 600.000000 600.000000 600.000000 600.000000 600.0 \n", | |
| "mean 715.301667 2.471633 17.616517 0.109882 1.0 \n", | |
| "std 405.232698 1.450937 28.350360 0.037981 0.0 \n", | |
| "min 6.000000 0.000000 -19.850000 0.023000 1.0 \n", | |
| "25% 361.750000 1.207500 -10.385000 0.082750 1.0 \n", | |
| "50% 734.500000 2.440000 17.520000 0.111000 1.0 \n", | |
| "75% 1070.250000 3.752500 45.835000 0.138250 1.0 \n", | |
| "max 1399.000000 4.990000 54.790000 0.195000 1.0 \n", | |
| "\n", | |
| " log_nitrate \n", | |
| "count 600.000000 \n", | |
| "mean -2.280787 \n", | |
| "std 0.405542 \n", | |
| "min -3.772261 \n", | |
| "25% -2.491945 \n", | |
| "50% -2.198225 \n", | |
| "75% -1.978697 \n", | |
| "max -1.634756 " | |
| ] | |
| }, | |
| "execution_count": 21, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df_extreme.describe()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 22, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<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>Unnamed: 0</th>\n", | |
| " <th>feed</th>\n", | |
| " <th>temperature</th>\n", | |
| " <th>nitrate</th>\n", | |
| " <th>extreme_temp</th>\n", | |
| " <th>log_nitrate</th>\n", | |
| " </tr>\n", | |
| " </thead>\n", | |
| " <tbody>\n", | |
| " <tr>\n", | |
| " <th>count</th>\n", | |
| " <td>800.000000</td>\n", | |
| " <td>800.000000</td>\n", | |
| " <td>800.000000</td>\n", | |
| " <td>800.000000</td>\n", | |
| " <td>800.0</td>\n", | |
| " <td>800.000000</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>mean</th>\n", | |
| " <td>687.648750</td>\n", | |
| " <td>7.375950</td>\n", | |
| " <td>17.280875</td>\n", | |
| " <td>0.057599</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>-3.026234</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>std</th>\n", | |
| " <td>403.427838</td>\n", | |
| " <td>1.433464</td>\n", | |
| " <td>9.879207</td>\n", | |
| " <td>0.028699</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>0.660617</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>min</th>\n", | |
| " <td>0.000000</td>\n", | |
| " <td>5.000000</td>\n", | |
| " <td>1.000000</td>\n", | |
| " <td>0.003000</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>-5.809143</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>25%</th>\n", | |
| " <td>344.750000</td>\n", | |
| " <td>6.117500</td>\n", | |
| " <td>8.675000</td>\n", | |
| " <td>0.034000</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>-3.381395</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>50%</th>\n", | |
| " <td>679.500000</td>\n", | |
| " <td>7.410000</td>\n", | |
| " <td>16.965000</td>\n", | |
| " <td>0.056000</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>-2.882404</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>75%</th>\n", | |
| " <td>1030.750000</td>\n", | |
| " <td>8.590000</td>\n", | |
| " <td>26.312500</td>\n", | |
| " <td>0.081000</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>-2.513306</td>\n", | |
| " </tr>\n", | |
| " <tr>\n", | |
| " <th>max</th>\n", | |
| " <td>1397.000000</td>\n", | |
| " <td>9.970000</td>\n", | |
| " <td>33.940000</td>\n", | |
| " <td>0.116000</td>\n", | |
| " <td>0.0</td>\n", | |
| " <td>-2.154165</td>\n", | |
| " </tr>\n", | |
| " </tbody>\n", | |
| "</table>\n", | |
| "</div>" | |
| ], | |
| "text/plain": [ | |
| " Unnamed: 0 feed temperature nitrate extreme_temp \\\n", | |
| "count 800.000000 800.000000 800.000000 800.000000 800.0 \n", | |
| "mean 687.648750 7.375950 17.280875 0.057599 0.0 \n", | |
| "std 403.427838 1.433464 9.879207 0.028699 0.0 \n", | |
| "min 0.000000 5.000000 1.000000 0.003000 0.0 \n", | |
| "25% 344.750000 6.117500 8.675000 0.034000 0.0 \n", | |
| "50% 679.500000 7.410000 16.965000 0.056000 0.0 \n", | |
| "75% 1030.750000 8.590000 26.312500 0.081000 0.0 \n", | |
| "max 1397.000000 9.970000 33.940000 0.116000 0.0 \n", | |
| "\n", | |
| " log_nitrate \n", | |
| "count 800.000000 \n", | |
| "mean -3.026234 \n", | |
| "std 0.660617 \n", | |
| "min -5.809143 \n", | |
| "25% -3.381395 \n", | |
| "50% -2.882404 \n", | |
| "75% -2.513306 \n", | |
| "max -2.154165 " | |
| ] | |
| }, | |
| "execution_count": 22, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df_regular.describe()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Analyze Results\n", | |
| "Looking at the feed column of each dataset description:\n", | |
| "- The feed production at non-extreme temperatures is between 5 and 10\n", | |
| "- The feed production at extreme temperatures is between 0 and 5\n", | |
| "\n", | |
| "So, it makes sense for the residuals to have two clusters of fitted values! Clearly, one cluster produces a greater amount -- in fact twice as much.\n", | |
| "\n", | |
| "From, this we can now make a **data-backed recommendation**:\n", | |
| "- Make sure all bioreactors operate between the temperatures of 0 and 34 to optimize algae feed production. This will double production compared to running at extreme temperatures." | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Step 12: Summarize the Client Deliverable!" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Model Deliverable\n", | |
| "\n", | |
| "The client can use this model to predict algae feed production in their bioreactors." | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 23, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/html": [ | |
| "<table class=\"simpletable\">\n", | |
| "<caption>OLS Regression Results</caption>\n", | |
| "<tr>\n", | |
| " <th>Dep. Variable:</th> <td>feed</td> <th> R-squared: </th> <td> 0.936</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Model:</th> <td>OLS</td> <th> Adj. R-squared: </th> <td> 0.936</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Method:</th> <td>Least Squares</td> <th> F-statistic: </th> <td> 6811.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Date:</th> <td>Sat, 21 Mar 2020</td> <th> Prob (F-statistic):</th> <td> 0.00</td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Time:</th> <td>17:37:48</td> <th> Log-Likelihood: </th> <td> -2566.4</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>No. Observations:</th> <td> 1400</td> <th> AIC: </th> <td> 5139.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Residuals:</th> <td> 1397</td> <th> BIC: </th> <td> 5155.</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Df Model:</th> <td> 3</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Covariance Type:</th> <td>nonrobust</td> <th> </th> <td> </td> \n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <td></td> <th>coef</th> <th>std err</th> <th>t</th> <th>P>|t|</th> <th>[0.025</th> <th>0.975]</th> \n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>nitrate</th> <td> 30.4056</td> <td> 0.971</td> <td> 31.325</td> <td> 0.000</td> <td> 28.501</td> <td> 32.310</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>extreme_temp</th> <td> -5.0676</td> <td> 0.104</td> <td> -48.839</td> <td> 0.000</td> <td> -5.271</td> <td> -4.864</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>log_nitrate</th> <td> -1.8407</td> <td> 0.022</td> <td> -81.815</td> <td> 0.000</td> <td> -1.885</td> <td> -1.797</td>\n", | |
| "</tr>\n", | |
| "</table>\n", | |
| "<table class=\"simpletable\">\n", | |
| "<tr>\n", | |
| " <th>Omnibus:</th> <td>179.156</td> <th> Durbin-Watson: </th> <td> 1.971</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Prob(Omnibus):</th> <td> 0.000</td> <th> Jarque-Bera (JB): </th> <td> 47.048</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Skew:</th> <td>-0.049</td> <th> Prob(JB): </th> <td>6.08e-11</td>\n", | |
| "</tr>\n", | |
| "<tr>\n", | |
| " <th>Kurtosis:</th> <td> 2.107</td> <th> Cond. No. </th> <td> 67.7</td>\n", | |
| "</tr>\n", | |
| "</table>" | |
| ], | |
| "text/plain": [ | |
| "<class 'statsmodels.iolib.summary.Summary'>\n", | |
| "\"\"\"\n", | |
| " OLS Regression Results \n", | |
| "==============================================================================\n", | |
| "Dep. Variable: feed R-squared: 0.936\n", | |
| "Model: OLS Adj. R-squared: 0.936\n", | |
| "Method: Least Squares F-statistic: 6811.\n", | |
| "Date: Sat, 21 Mar 2020 Prob (F-statistic): 0.00\n", | |
| "Time: 17:37:48 Log-Likelihood: -2566.4\n", | |
| "No. Observations: 1400 AIC: 5139.\n", | |
| "Df Residuals: 1397 BIC: 5155.\n", | |
| "Df Model: 3 \n", | |
| "Covariance Type: nonrobust \n", | |
| "================================================================================\n", | |
| " coef std err t P>|t| [0.025 0.975]\n", | |
| "--------------------------------------------------------------------------------\n", | |
| "nitrate 30.4056 0.971 31.325 0.000 28.501 32.310\n", | |
| "extreme_temp -5.0676 0.104 -48.839 0.000 -5.271 -4.864\n", | |
| "log_nitrate -1.8407 0.022 -81.815 0.000 -1.885 -1.797\n", | |
| "==============================================================================\n", | |
| "Omnibus: 179.156 Durbin-Watson: 1.971\n", | |
| "Prob(Omnibus): 0.000 Jarque-Bera (JB): 47.048\n", | |
| "Skew: -0.049 Prob(JB): 6.08e-11\n", | |
| "Kurtosis: 2.107 Cond. No. 67.7\n", | |
| "==============================================================================\n", | |
| "\n", | |
| "Warnings:\n", | |
| "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n", | |
| "\"\"\"" | |
| ] | |
| }, | |
| "execution_count": 23, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "df = pd.read_csv('FeedProduction.csv')\n", | |
| "df['extreme_temp'] = [1 if x <= 0 or x >= 35 else 0 for x in df['temperature']]\n", | |
| "df['log_nitrate'] = np.log(df['nitrate'])\n", | |
| "X = df[['nitrate', 'extreme_temp', 'log_nitrate']]\n", | |
| "y = df['feed']\n", | |
| "model = sm.OLS(y, X).fit()\n", | |
| "model.summary()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "### Actionable Insight Deliverable\n", | |
| "\n", | |
| "Keep bioreactor temperatures between 0 and 34. This will **double production** compared to running at extreme temperatures." | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "# Conclusion\n", | |
| "\n", | |
| "In this case, we learned how to:\n", | |
| "- Run a linear regression in Python\n", | |
| "- Interpret linear regression results\n", | |
| "- Identify needs for new features and create them\n", | |
| "- Interpret residual plots\n", | |
| "- Produce an actionable client deliverable\n", | |
| "- **Iterate on a model**\n", | |
| "\n", | |
| "The last bulletpoint is the key takeaway. The final deliverable is very concise, but it took a lot of iteration and analysis to get there. Although the final deliverable is concise, it actually has huge ramifications for FarmX. \n", | |
| "- FarmX can now predict algae feed production using easy-to-measure variables (no new fancy sensors or staffing!)\n", | |
| "- FarmX has insight on how to optimize feed production up to twice as much as before in some cases\n", | |
| "- **Productivity and profitability can increase without any new costs**\n", | |
| "\n", | |
| "And because the final deliverable is concise, it is easy to communicate and apply. A deliverable that cannot be understood is a deliverable that will never be used, no matter how well it actually works.\n", | |
| "\n", | |
| "Additionally, consider how little we actually coded. Coding is a tool for data science and data analytics, but at its core data science is an understanding of math (which is why it is important to learn the theory!). It is easy to run a model -- even something as complex as a neural net -- but it is more important to understand how to do it right and when to apply certain methods. As you've hopefully seen, something as fundamental as linear regression can be quite extensive and powerful.\n" | |
| ] | |
| } | |
| ], | |
| "metadata": { | |
| "kernelspec": { | |
| "display_name": "Python 3", | |
| "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": "3.6.3" | |
| } | |
| }, | |
| "nbformat": 4, | |
| "nbformat_minor": 2 | |
| } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment