Skip to content

Instantly share code, notes, and snippets.

@naviarh

naviarh/probe2.ipynb

Created August 1, 2018 19:10
Show Gist options
  • Save naviarh/7b2be5fd8abf2833729fd3bf6e8be1e5 to your computer and use it in GitHub Desktop.
Save naviarh/7b2be5fd8abf2833729fd3bf6e8be1e5 to your computer and use it in GitHub Desktop.
Download ZIP
probe2.ipynb
Display the source blob
Display the rendered blob
Raw
probe2.ipynb
{
"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"colab": {
"name": "probe2.ipynb",
"version": "0.3.2",
"provenance": [],
"collapsed_sections": [],
"include_colab_link": true
},
"kernelspec": {
"name": "python3",
"display_name": "Python 3"
},
"accelerator": "GPU"
},
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "view-in-github",
"colab_type": "text"
},
"source": [
"[View in Colaboratory](https://colab.research.google.com/gist/naviarh/7b2be5fd8abf2833729fd3bf6e8be1e5/probe2.ipynb)"
]
},
{
"metadata": {
"id": "pRjLc4Z8NJK_",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"Алгоритм:\n",
"1. Находим все возможные полезные признаки;\n",
"2. Разделяем признаки на количественные и категориальные;\n",
"3. Формируем датафрейм признаков;\n",
"4. Создаём валидационную выборку;\n",
"5. Объект данных (ColumnarModelData.from_data_frame())\n",
"6. Создаём массив размеров матриц эмбеддинга\n",
"7. Создаём алгоритм обучения (моделль нейросети)\n",
"8. Обучаем модель\n"
]
},
{
"metadata": {
"id": "DrEiybPJLfEh",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"### Настройка Google Colaboratory"
]
},
{
"metadata": {
"id": "oLQi1xH4Lcv5",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"!pip install --upgrade pip\n",
"!pip3 install fastai"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "slMuW4iLDZkn",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"### Настройка платформы"
]
},
{
"metadata": {
"id": "a2D3n9FcMFNz",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "a5d91a01-1c1c-4276-d4b7-3125a9c74040"
},
"cell_type": "code",
"source": [
"# Put these at the top of every notebook, to get automatic reloading and inline plotting\n",
"%reload_ext autoreload\n",
"%autoreload 2\n",
"%matplotlib inline\n",
"!pwd"
],
"execution_count": 159,
"outputs": [
{
"output_type": "stream",
"text": [
"/content\r\n"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "HrPDR7_0uEWN",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"#!/usr/bin/env python3\n",
"# -*- coding: utf-8 -*-"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "rpY3XcKXXg_t",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 59
},
"outputId": "de9ae2d0-9206-43e4-c5c2-24b0c66c2322"
},
"cell_type": "code",
"source": [
"import subprocess, os\n",
"os.uname()"
],
"execution_count": 161,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"posix.uname_result(sysname='Linux', nodename='4feb6a706292', release='4.14.33+', version='#1 SMP Wed Jun 20 01:15:52 PDT 2018', machine='x86_64')"
]
},
"metadata": {
"tags": []
},
"execution_count": 161
}
]
},
{
"metadata": {
"id": "2BjQb7-MXjow",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 102
},
"outputId": "de8d0064-6d5e-4d27-d559-fc24bd3bda3b"
},
"cell_type": "code",
"source": [
"from fastai.structured import *\n",
"from fastai.column_data import *\n",
"!pip3 show fastai torch | grep Name -A 1"
],
"execution_count": 162,
"outputs": [
{
"output_type": "stream",
"text": [
"Name: fastai\r\n",
"Version: 0.7.0\r\n",
"--\r\n",
"Name: torch\r\n",
"Version: 0.3.1\r\n"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "mmGfSBAAYR6q",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"torch.cuda.set_device(0)\n",
"torch.cuda.get_device_name(0)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "hByGgSB2XxXj",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"#from fastai.imports import *\n",
"#from fastai.transforms import *\n",
"#from fastai.conv_learner import *\n",
"#from fastai.model import *\n",
"#from fastai.dataset import *\n",
"#from fastai.sgdr import *\n",
"#from fastai.plots import *\n",
"#!pip3 show fastai torch | grep Name -A 1"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "N8A6-IJlZhwm",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Устанавливаем формат показа массивов numpy\n",
"np.set_printoptions(threshold=50, edgeitems=20)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "1B9wSpR6BvDe",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Создадим путь к каталогу PATH\n",
"PATH='btc/'\n",
"os.makedirs(PATH, exist_ok=True)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "8F5gx4M1SN8o",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"from google.colab import files"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "EZZkvdu-YCXO",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"### Подготовка данных"
]
},
{
"metadata": {
"id": "NQIjwvD2aY_3",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Загрузка исторических данных по дням для BTC-USD с 1 января по 29 июля 2018\n",
"# Скачать здесь: https://drive.google.com/open?id=1VI31_Gqt2ly83v4kxsN0_FnDPSeplF0l\n",
"# \n",
"# Сами данные на сайте: https://finance.yahoo.com/quote/BTC-USD/history?p=BTC-USD\n",
"uploaded = files.upload()"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "4glnYHfWbLJs",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Видим нужный файл BTC-USD.csv\n",
"!ls"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "dUJjNutYSkcs",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Чтение данных\n",
"data = pd.read_csv('BTC-USD.csv', error_bad_lines=False)[::1]\n",
"data.shape"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "6p_ieeGaPUuk",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Здесь видим график цены по дням\n",
"price = data.loc[:, 'Adj Close'].tolist()\n",
"plt.figure(figsize=(12,5))\n",
"plt.plot(price)\n",
"plt.show()"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "aqmH49FWPrmR",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Таблица цен и объёмов по дням\n",
"data.head()"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "DpmZZfCY6Zvk",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Массив дат\n",
"date = data['Date']\n",
"date.head(5)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "Cu7vFv7q6xaf",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Расширяем информацию о датах (всё что знает pandas о днях)\n",
"add_datepart(data, \"Date\", drop=False)\n",
"data.head(1)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "Csgve0GQ1NtJ",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Удалим столбцы: дату, день месяца и день года\n",
"#data.drop('Adj Close', axis=1, inplace=True)\n",
"data.drop('Date', axis=1, inplace=True)\n",
"data.drop('Day', axis=1, inplace=True)\n",
"data.drop('Dayofyear', axis=1, inplace=True)\n",
"data.head(2)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "q9F0cgtVMUuS",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Переводим год в натуральный ряд\n",
"data['Year'] -= np.min(data['Year'])\n",
"data.head(1)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "YroW8w9N2T05",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Заменяем булевые значения бинарными\n",
"data.Is_month_end = data.Is_month_end.astype('int')\n",
"data.Is_month_start = data.Is_month_start.astype('int')\n",
"data.Is_quarter_end = data.Is_quarter_end.astype('int')\n",
"data.Is_quarter_start = data.Is_quarter_start.astype('int')\n",
"data.Is_year_end = data.Is_year_end.astype('int')\n",
"data.Is_year_start = data.Is_year_start.astype('int')\n",
"data.head(8)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "iyvGI0kyjxzy",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"len(data)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "qhOUgbJ54rFd",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Разделим данные на тестовый и обучающий набор\n",
"len_test = 100 # последние 100 строк\n",
"\n",
"test = data[-len_test:]\n",
"data = data[:len(data)-len(test)]\n",
"\n",
"date_test = date[-len_test:]\n",
"date_data = date[:len(date)-len(date_test)]\n",
"\n",
"len(test), len(date_test)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "ygIvBOBiSPEt",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Количество обучающих записей и количество признаков (фич)\n",
"data.shape"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "c-El6tFkpibX",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Делаем столбец даты индексом\n",
"#data.set_index('Date', inplace=True)\n",
"#data.head(2)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "R98nw8IXqRLj",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Просмотр фич (признаков)\n",
"data.columns"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "P60NrFN4qZxa",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Категориальные данные\n",
"categ_vars = ['Year', 'Month', 'Week', 'Dayofweek']\n",
"\n",
"# Бинарные данные\n",
"binary_vars = ['Is_month_end', 'Is_month_start', 'Is_quarter_end',\n",
" 'Is_quarter_start', 'Is_year_end', 'Is_year_start']\n",
"\n",
"# Количественные данные\n",
"contin_vars = ['Open', 'High', 'Low', 'Close', 'Volume', 'Elapsed']\n",
"\n",
"len(categ_vars)+len(binary_vars)+len(contin_vars), len(categ_vars), len(binary_vars), len(contin_vars)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "Jbv75qnC6yxQ",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Все категориальные данные (нужно для разделения обучающих и тестовых данных)\n",
"cat_bin = categ_vars+binary_vars\n",
"cat_bin"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "xeIA4A5IDr1b",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Разделим обучающие данные для нейросети\n",
"df, y, nas, mapper = proc_df(data, 'Adj Close', ignore_flds=cat_bin, do_scale=True,) # 'Adj Close' - на входит в df\n",
"y = np.log(y+1) # Логарифмируем целевую переменную, увеличенную на 1 для неотрицательности (потом брать экспоненту: np.exp(y)-1)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "ttXN0TqkEjMm",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Сместим курсы на день назад для определения предсказаний\n",
"df = df[:-1] # уберём в связи с этим последнюю выборку\n",
"date_data = date_data[:-1]\n",
"y = np.delete(y, (0))\n",
"y"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "lTGGHyTpegjD",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"np.min(y)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "ua5UqGYvfl_K",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"np.max(y)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "f35ixRvfdvMQ",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"len(df)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "RU7n9_VWFN38",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Разделим тестовые данные для нейросети\n",
"df_test, y_test, nas, mapper = proc_df(test, 'Adj Close', ignore_flds=cat_bin, do_scale=True, mapper=mapper, na_dict=nas) # 'Adj Close' - на входит в df\n",
"y_test = np.log(y_test) # Логарифмируем целевую переменную (потом брать экспоненту: np.exp(y_test))"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "rcfyoPUJFovW",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Сместим тестовые курсы на день назад для определения предсказаний\n",
"df_test = df_test[:-1] # уберём в связи с этим последнюю выборку\n",
"date_test = date_test[:-1]\n",
"y_test = np.delete(y_test, (0))\n",
"y_test"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "RpLEC3vSQR0h",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"df_test.head(2)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "ZoxkbZhtGCsV",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Итоговый обучающий датафрейм\n",
"df.head()"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "hvqkFUPyGG4F",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Проверим цены\n",
"np.exp(y)-1"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "ehmWI9zLpbpk",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Сохраняем данные\n",
"datas = df\n",
"datas['y'] = y\n",
"datas['date'] = date_data\n",
"datas.to_csv(f'{PATH}datas.csv')\n",
"\n",
"tests = df_test\n",
"tests['y_test'] = y_test\n",
"tests['date_test'] = date_test\n",
"tests.to_csv(f'{PATH}tests.csv')\n",
"\n",
"!ls {PATH}"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "kLYfUNWAzDaV",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Скачать обучающий набор\n",
"files.download(f'{PATH}datas.csv')"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "_CTwH_h_5ufo",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Скачать тестовый набор\n",
"files.download(f'{PATH}tests.csv')"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "NAM44HrYtfc1",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Cозранение и загрузка numpy массивов\n",
"'''\n",
"f = open(f'{PATH}data','wb')\n",
"np.save(f,data)\n",
"f.close()\n",
"\n",
"f = open(f'{PATH}data','rb')\n",
"data = np.load(f)\n",
"f.close()\n",
"\n",
"!head -n 3 '{PATH}data'\n",
"'''"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "9lg_qMlptl3s",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Быстрые сохранение и загрузка pandas\n",
"#data.to_feather(f'{PATH}data')\n",
"#data = pd.read_feather(f'{PATH}data')\n",
"#!head -n 3 '{PATH}data'"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "lFw9pKf-tFOX",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"### Создание модели"
]
},
{
"metadata": {
"id": "yyDsevZsOGZT",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "85e97efa-64c2-4817-b94c-977550174020"
},
"cell_type": "code",
"source": [
"# Проверить наличие данных\n",
"!ls {PATH}"
],
"execution_count": 168,
"outputs": [
{
"output_type": "stream",
"text": [
"datas.csv models tests.csv tmp\r\n"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "RKLMANJFV7o9",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"Обучающий и тестовый наборы данных можно скачать с Google Drive:\n",
"\n",
"https://drive.google.com/open?id=1VciXZwYf28cPMcpAmXEP4wvmjHqVBvvM\n",
"\n",
"https://drive.google.com/open?id=1U_JYQwjql1AyVjxAfBt0u5xMphGeA6ft"
]
},
{
"metadata": {
"id": "HzvGuyu1Aoqz",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Загрузить данные обучающий набор datas.csv\n",
"uploaded = files.upload()\n",
"\n",
"# Загрузить данные тестовый набор tests.csv\n",
"uploaded = files.upload()"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "1q_bVNk7Ntyh",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Скопировать данные в рабочий каталог\n",
"!mv datas.csv '{PATH}'\n",
"!mv tests.csv '{PATH}'\n",
"!ls {PATH}"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "n8l8TSbMpRAc",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 119
},
"outputId": "bb9c4094-e603-4cd5-fc15-b003d2c5345f"
},
"cell_type": "code",
"source": [
"# Чтение данных\n",
"df = pd.read_csv(f'{PATH}datas.csv')\n",
"df_test = pd.read_csv(f'{PATH}tests.csv')\n",
"\n",
"y = df['y']\n",
"date_data = df['date']\n",
"y_test = df_test['y_test']\n",
"date_test = df_test['date_test']\n",
"\n",
"df.drop('y', axis=1, inplace=True)\n",
"df.drop('date', axis=1, inplace=True)\n",
"df_test.drop('y_test', axis=1, inplace=True)\n",
"df_test.drop('date_test', axis=1, inplace=True)\n",
"\n",
"date_data[:5]"
],
"execution_count": 264,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0 2010-07-16\n",
"1 2010-07-17\n",
"2 2010-07-18\n",
"3 2010-07-19\n",
"4 2010-07-20\n",
"Name: date, dtype: object"
]
},
"metadata": {
"tags": []
},
"execution_count": 264
}
]
},
{
"metadata": {
"id": "CWM04b4wH7Gl",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 246
},
"outputId": "5168ffe8-de1e-47a7-90f7-60a2b0402318"
},
"cell_type": "code",
"source": [
"# Обучающий датафрейм\n",
"df.head()"
],
"execution_count": 265,
"outputs": [
{
"output_type": "execute_result",
"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>Year</th>\n",
" <th>Month</th>\n",
" <th>Week</th>\n",
" <th>Dayofweek</th>\n",
" <th>Is_month_end</th>\n",
" <th>Is_month_start</th>\n",
" <th>Is_quarter_end</th>\n",
" <th>Is_quarter_start</th>\n",
" <th>Is_year_end</th>\n",
" <th>Is_year_start</th>\n",
" <th>Open</th>\n",
" <th>High</th>\n",
" <th>Low</th>\n",
" <th>Close</th>\n",
" <th>Volume</th>\n",
" <th>Elapsed</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>7</td>\n",
" <td>28</td>\n",
" <td>4</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>-0.412458</td>\n",
" <td>-0.410293</td>\n",
" <td>-0.417710</td>\n",
" <td>-0.413042</td>\n",
" <td>-0.298214</td>\n",
" <td>-1.731912</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>7</td>\n",
" <td>28</td>\n",
" <td>5</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>-0.412458</td>\n",
" <td>-0.410280</td>\n",
" <td>-0.417706</td>\n",
" <td>-0.413029</td>\n",
" <td>-0.298214</td>\n",
" <td>-1.730689</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2</td>\n",
" <td>0</td>\n",
" <td>7</td>\n",
" <td>28</td>\n",
" <td>6</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>-0.412445</td>\n",
" <td>-0.410277</td>\n",
" <td>-0.417699</td>\n",
" <td>-0.413031</td>\n",
" <td>-0.298213</td>\n",
" <td>-1.729465</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>3</td>\n",
" <td>0</td>\n",
" <td>7</td>\n",
" <td>29</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>-0.412446</td>\n",
" <td>-0.410282</td>\n",
" <td>-0.417700</td>\n",
" <td>-0.413033</td>\n",
" <td>-0.298214</td>\n",
" <td>-1.728242</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>4</td>\n",
" <td>0</td>\n",
" <td>7</td>\n",
" <td>29</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>-0.412449</td>\n",
" <td>-0.410283</td>\n",
" <td>-0.417704</td>\n",
" <td>-0.413031</td>\n",
" <td>-0.298214</td>\n",
" <td>-1.727018</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Unnamed: 0 Year Month Week Dayofweek Is_month_end Is_month_start \\\n",
"0 0 0 7 28 4 0 0 \n",
"1 1 0 7 28 5 0 0 \n",
"2 2 0 7 28 6 0 0 \n",
"3 3 0 7 29 0 0 0 \n",
"4 4 0 7 29 1 0 0 \n",
"\n",
" Is_quarter_end Is_quarter_start Is_year_end Is_year_start Open \\\n",
"0 0 0 0 0 -0.412458 \n",
"1 0 0 0 0 -0.412458 \n",
"2 0 0 0 0 -0.412445 \n",
"3 0 0 0 0 -0.412446 \n",
"4 0 0 0 0 -0.412449 \n",
"\n",
" High Low Close Volume Elapsed \n",
"0 -0.410293 -0.417710 -0.413042 -0.298214 -1.731912 \n",
"1 -0.410280 -0.417706 -0.413029 -0.298214 -1.730689 \n",
"2 -0.410277 -0.417699 -0.413031 -0.298213 -1.729465 \n",
"3 -0.410282 -0.417700 -0.413033 -0.298214 -1.728242 \n",
"4 -0.410283 -0.417704 -0.413031 -0.298214 -1.727018 "
]
},
"metadata": {
"tags": []
},
"execution_count": 265
}
]
},
{
"metadata": {
"id": "nmZQKmozIAdl",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 119
},
"outputId": "92353323-abbc-4598-b6d3-addb1bd83d3a"
},
"cell_type": "code",
"source": [
"# Обучающие цены\n",
"np.exp(y[-5:])-1"
],
"execution_count": 266,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"2825 7927.729980\n",
"2826 7899.109863\n",
"2827 8022.509766\n",
"2828 8376.730469\n",
"2829 8079.770020\n",
"Name: y, dtype: float64"
]
},
"metadata": {
"tags": []
},
"execution_count": 266
}
]
},
{
"metadata": {
"id": "HGNZ5ubXRU1m",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Удалить лишний столбец, если есть (\"Unnamed: 0\")\n",
"df.drop('Unnamed: 0', axis=1, inplace=True)\n",
"df_test.drop('Unnamed: 0', axis=1, inplace=True)"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "lNaVd1HyG1rq",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 85
},
"outputId": "ed791ee1-b17d-4ca7-d9e0-6c313020890b"
},
"cell_type": "code",
"source": [
"# Просмотр фич (признаков)\n",
"df.columns"
],
"execution_count": 268,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"Index(['Year', 'Month', 'Week', 'Dayofweek', 'Is_month_end', 'Is_month_start',\n",
" 'Is_quarter_end', 'Is_quarter_start', 'Is_year_end', 'Is_year_start',\n",
" 'Open', 'High', 'Low', 'Close', 'Volume', 'Elapsed'],\n",
" dtype='object')"
]
},
"metadata": {
"tags": []
},
"execution_count": 268
}
]
},
{
"metadata": {
"id": "_9audFp9F7Ol",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "2b3e4f8e-8f5b-4cd9-beb9-2ade9b402681"
},
"cell_type": "code",
"source": [
"# Категориальные данные\n",
"categ_vars = ['Year', 'Month', 'Week', 'Dayofweek']\n",
"\n",
"# Бинарные данные\n",
"binary_vars = ['Is_month_end', 'Is_month_start', 'Is_quarter_end',\n",
" 'Is_quarter_start', 'Is_year_end', 'Is_year_start']\n",
"\n",
"# Количественные данные\n",
"contin_vars = ['Open', 'High', 'Low', 'Close', 'Volume', 'Elapsed']\n",
"\n",
"len(categ_vars)+len(binary_vars)+len(contin_vars), len(categ_vars), len(binary_vars), len(contin_vars)"
],
"execution_count": 269,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(16, 4, 6, 6)"
]
},
"metadata": {
"tags": []
},
"execution_count": 269
}
]
},
{
"metadata": {
"id": "zRGWmM-cGGtr",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 187
},
"outputId": "422e6493-9b02-4bc1-8f35-bed7eaf4cb2d"
},
"cell_type": "code",
"source": [
"# Все категориальные данные (нужно для матриц эмбеддинга)\n",
"cat_bin = categ_vars+binary_vars\n",
"cat_bin"
],
"execution_count": 270,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"['Year',\n",
" 'Month',\n",
" 'Week',\n",
" 'Dayofweek',\n",
" 'Is_month_end',\n",
" 'Is_month_start',\n",
" 'Is_quarter_end',\n",
" 'Is_quarter_start',\n",
" 'Is_year_end',\n",
" 'Is_year_start']"
]
},
"metadata": {
"tags": []
},
"execution_count": 270
}
]
},
{
"metadata": {
"id": "JmLa-KnmGWND",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "e223040c-7dd7-4d2d-cea5-7fcfd2f97c1b"
},
"cell_type": "code",
"source": [
"# Размеры категорий (+1 для неизвестного нейронке значения признака)\n",
"categ_sz = [(c, len(np.unique(df[c]))+1) for c in categ_vars]\n",
"categ_sz # мощностя категорий"
],
"execution_count": 271,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[('Year', 10), ('Month', 13), ('Week', 54), ('Dayofweek', 8)]"
]
},
"metadata": {
"tags": []
},
"execution_count": 271
}
]
},
{
"metadata": {
"id": "k3jyWh18Gf31",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 119
},
"outputId": "c4ca626c-6adb-4757-809f-cf1c913049ab"
},
"cell_type": "code",
"source": [
"# Размеры бинарных данных\n",
"binary_sz = [(c, 3) for c in binary_vars]\n",
"binary_sz # мощностя бинаров = 3"
],
"execution_count": 272,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[('Is_month_end', 3),\n",
" ('Is_month_start', 3),\n",
" ('Is_quarter_end', 3),\n",
" ('Is_quarter_start', 3),\n",
" ('Is_year_end', 3),\n",
" ('Is_year_start', 3)]"
]
},
"metadata": {
"tags": []
},
"execution_count": 272
}
]
},
{
"metadata": {
"id": "FfxHyhkxGl6i",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 187
},
"outputId": "8a56ebd6-e8d3-4984-9efe-c5a1eb4bc9ec"
},
"cell_type": "code",
"source": [
"# Размеры матриц эмбэддинга\n",
"emb_szs = [(c, min(50, (c+1)//2)) for _,c in categ_sz]\n",
"emb_szs += [(c, 2) for _,c in binary_sz]\n",
"emb_szs"
],
"execution_count": 273,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[(10, 5),\n",
" (13, 7),\n",
" (54, 27),\n",
" (8, 4),\n",
" (3, 2),\n",
" (3, 2),\n",
" (3, 2),\n",
" (3, 2),\n",
" (3, 2),\n",
" (3, 2)]"
]
},
"metadata": {
"tags": []
},
"execution_count": 273
}
]
},
{
"metadata": {
"id": "hwOg45CWPSvN",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Формируем случайные индексы валидационной выборки\n",
"val_ratio = 0.2 # 20% от обучающей выборки\n",
"\n",
"len_val = int(len(df)*val_ratio)-1\n",
"print(len_val) # размер валидационной выборки\n",
"\n",
"val_idx = np.random.choice(len(df), size=len_val, replace=False)\n",
"val_idx # индексы валидационной выборки"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "0z1DUySSPcdB",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 68
},
"outputId": "ebf31ab9-0df0-439e-9cb1-9eed6eebc140"
},
"cell_type": "code",
"source": [
"# Формируем индексы валидационной выборки из функции fastai\n",
"val_idx = get_cv_idxs(len(df)-1)\n",
"val_idx"
],
"execution_count": 274,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([1025, 1295, 2671, 1583, 196, 1832, 322, 1298, 2329, 1452, 1565, 324, 554, 1543, 817, 2288, 2010,\n",
" 1270, 803, 1448, ..., 1421, 1577, 857, 2201, 1211, 565, 2731, 1330, 788, 1501, 1230, 435, 691,\n",
" 237, 789, 109, 1849, 163, 1878, 252])"
]
},
"metadata": {
"tags": []
},
"execution_count": 274
}
]
},
{
"metadata": {
"id": "KGr6wMXUPk1J",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "e0f75860-40f5-4856-f82d-933a2969f692"
},
"cell_type": "code",
"source": [
"print(len(val_idx)) # размер валидационной выборки"
],
"execution_count": 275,
"outputs": [
{
"output_type": "stream",
"text": [
"565\n"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "Q6IusIHWHuXX",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Получаем диапазон цен для нейросети\n",
"y_range = (0, np.max(y)*1.5)\n",
"y_range_test = y_range\n",
"y_range"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "t2uNUK7fJots",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "d282dfc4-22be-4aa2-c996-92122e89f5e7"
},
"cell_type": "code",
"source": [
"# Создаём метрику оценки качества модели\n",
"\n",
"# 1 - |1 - y_pred/targ|\n",
"\n",
"def inv_y(a): return np.exp(a) # восстановление цены курсов\n",
"\n",
"def metrics(y_pred, targ):\n",
" result = 1 - np.absolute(1 - y_pred/targ)\n",
" return result.mean()\n",
"metrics"
],
"execution_count": 277,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"<function __main__.metrics>"
]
},
"metadata": {
"tags": []
},
"execution_count": 277
}
]
},
{
"metadata": {
"id": "G1k4cYtWTdk1",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "d976c826-fc43-4962-f318-dace6e1d8146"
},
"cell_type": "code",
"source": [
"# Данные модели для обучения\n",
"md = ColumnarModelData.from_data_frame(PATH, # каталог модели\n",
" val_idx, # валидационная выборка\n",
" df, # обучающий датафрейм\n",
" y.astype(np.float32), # целевая переменная\n",
" cat_flds=cat_bin, # категориальные признаки\n",
" bs=1, # размер минибатча\n",
" test_df=df_test) # тестоввая выбока\n",
"md"
],
"execution_count": 278,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"<fastai.column_data.ColumnarModelData at 0x7fc14554bf60>"
]
},
"metadata": {
"tags": []
},
"execution_count": 278
}
]
},
{
"metadata": {
"id": "mohMAOaXTdpl",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 510
},
"outputId": "f95b6b54-e3d4-4ecf-8ed8-6aee7f4df5bf"
},
"cell_type": "code",
"source": [
"# Модель нейронной сети\n",
"m = md.get_learner(emb_szs, # матрица эмбеддинга для категориальных признаков\n",
" len(contin_vars), # количество количественных признаков\n",
" 0.04, # дропаут для эмбеддинга\n",
" 1, # Количество чисел на выходе (мы предсказываем продажи)\n",
" [1000,500], # список количества активаций каждого слоя\n",
" [0.001,0.01])#, # список дропаутов для каждого слоя\n",
" #y_range=y_range) # желаемый диапазон значени целевой переменной\n",
"m "
],
"execution_count": 279,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"MixedInputModel(\n",
" (embs): ModuleList(\n",
" (0): Embedding(10, 5)\n",
" (1): Embedding(13, 7)\n",
" (2): Embedding(54, 27)\n",
" (3): Embedding(8, 4)\n",
" (4): Embedding(3, 2)\n",
" (5): Embedding(3, 2)\n",
" (6): Embedding(3, 2)\n",
" (7): Embedding(3, 2)\n",
" (8): Embedding(3, 2)\n",
" (9): Embedding(3, 2)\n",
" )\n",
" (lins): ModuleList(\n",
" (0): Linear(in_features=61, out_features=1000, bias=True)\n",
" (1): Linear(in_features=1000, out_features=500, bias=True)\n",
" )\n",
" (bns): ModuleList(\n",
" (0): BatchNorm1d(1000, eps=1e-05, momentum=0.1, affine=True)\n",
" (1): BatchNorm1d(500, eps=1e-05, momentum=0.1, affine=True)\n",
" )\n",
" (outp): Linear(in_features=500, out_features=1, bias=True)\n",
" (emb_drop): Dropout(p=0.04)\n",
" (drops): ModuleList(\n",
" (0): Dropout(p=0.001)\n",
" (1): Dropout(p=0.01)\n",
" )\n",
" (bn): BatchNorm1d(6, eps=1e-05, momentum=0.1, affine=True)\n",
")"
]
},
"metadata": {
"tags": []
},
"execution_count": 279
}
]
},
{
"metadata": {
"id": "XgrI9Oa3QngO",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"### Оптимизация модели"
]
},
{
"metadata": {
"id": "I5NPRbL1Tdgl",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 51
},
"outputId": "de202660-df2f-42f5-a8b4-d9a08cb5adc8"
},
"cell_type": "code",
"source": [
"# Вычисление скорости обучения\n",
"m.lr_find()"
],
"execution_count": 280,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "6ac5c569eb63450baaf0ae07ecb8b6d0",
"version_minor": 0,
"version_major": 2
},
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=1), HTML(value='')))"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
" 36%|███▋ | 822/2265 [00:15<00:26, 54.57it/s, loss=5.09]"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "gQJKANu-7i9_",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 283
},
"outputId": "a92748b7-65e0-4b98-ae62-54e32d4c72d4"
},
"cell_type": "code",
"source": [
"# График подбора скорости обучения\n",
"m.sched.plot(100)"
],
"execution_count": 281,
"outputs": [
{
"output_type": "display_data",
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKKWU8tOkoJRSyk+TglJKKb//DwEgtU8BnCdBAAAAAElFTkSuQmCC\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7fc144759780>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"metadata": {
"id": "emF6CvwbiaYA",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"Проводим обучение"
]
},
{
"metadata": {
"id": "LWeLAbqTJo6u",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Скорость обучениия\n",
"lr = 5e-3"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "dcqPNe_aSRqO",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 85
},
"outputId": "3ecfdc16-75cc-4e19-af13-a81d29b3fb49"
},
"cell_type": "code",
"source": [
"m.fit(lr, 1, metrics=[metrics])"
],
"execution_count": 283,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "f7bbba9680074f4ebee060467df39485",
"version_minor": 0,
"version_major": 2
},
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=1), HTML(value='')))"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
"epoch trn_loss val_loss metrics \n",
" 0 0.47801 0.37449 0.69971 \n"
],
"name": "stdout"
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[array([0.37449]), 0.6997098530288292]"
]
},
"metadata": {
"tags": []
},
"execution_count": 283
}
]
},
{
"metadata": {
"id": "S-c7s-WIoBBq",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 85
},
"outputId": "e4630c41-14ac-423f-8da7-b6ebec3eb959"
},
"cell_type": "code",
"source": [
"m.fit(lr, 1, cycle_len=1, metrics=[metrics])"
],
"execution_count": 284,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "c88e5005871b4f87b009c964dd90b179",
"version_minor": 0,
"version_major": 2
},
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=1), HTML(value='')))"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
"epoch trn_loss val_loss metrics \n",
" 0 0.142387 0.0864 0.899063 \n"
],
"name": "stdout"
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[array([0.0864]), 0.8990628219283787]"
]
},
"metadata": {
"tags": []
},
"execution_count": 284
}
]
},
{
"metadata": {
"id": "oI30zGYkse3e",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"lr = 1e-4"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "XuImpz-nbUlY",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 187
},
"outputId": "95481c32-eae0-4f44-a66f-44cf42490a2d"
},
"cell_type": "code",
"source": [
"m.fit(lr, 3, cycle_len=1, cycle_mult=2, metrics=[metrics])"
],
"execution_count": 286,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "41cb11926100480c9f081f41d8c77481",
"version_minor": 0,
"version_major": 2
},
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=7), HTML(value='')))"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
""
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
"epoch trn_loss val_loss metrics \n",
" 0 0.267485 0.073865 0.907383 \n",
" 25%|██▍ | 564/2265 [00:10<00:31, 53.87it/s, loss=0.0843]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 1 0.086927 0.058872 0.903964 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0934]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 2 0.152355 0.056512 0.918608 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.138]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 3 0.128719 0.048559 0.921147 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.117]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 4 0.070737 0.042772 0.915661 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0645]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 5 0.080302 0.037747 0.932512 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0802]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 6 0.070989 0.037578 0.932977 \n"
],
"name": "stdout"
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[array([0.03758]), 0.9329774954677683]"
]
},
"metadata": {
"tags": []
},
"execution_count": 286
}
]
},
{
"metadata": {
"id": "_4BuAvPWPoQL",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 279
},
"outputId": "98cd71c0-7b5c-4e52-c30b-a57839a9e841"
},
"cell_type": "code",
"source": [
"# График изменения скорости обучения\n",
"m.sched.plot_lr()"
],
"execution_count": 287,
"outputs": [
{
"output_type": "display_data",
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAZgAAAEGCAYAAABYV4NmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xl8lfWZ///Xfc7Jvu8JIYQkhA+7\nyKKgqAi41qUWrVbbunVTp2PbaWfsfDs62mVmfm3H1mmntVO7aFvrbrUqKoqioIKg7FwsIRAIkABZ\nScj+++O+gxETckjOyX2W6/l48Ai5z33f531OCNe5P9tt9fT0oJRSSgWax+0ASimlIpMWGKWUUkGh\nBUYppVRQaIFRSikVFFpglFJKBYXP7QChoLa2achD6TIyEqmrawlknIAI1VwQutk016kL1WyhmgtC\nN9tQcuXkpFgne1yvYIbJ5/O6HaFfoZoLQjeb5jp1oZotVHNB6GYLRi4tMEoppYJCC4xSSqmg0AKj\nlFIqKLTAKKWUCgotMEoppYIiqMOUjTH3A3OAHuBOEVnd57FFwI+ALuBFEfn+QMcYY4qARwAvsB/4\ngoi0GWMygEeBZhG52jk+BvgDUOyc+2YRqQjm61RKKfVJQbuCMcacB5SLyFzgVuCBE3Z5AFgMnA1c\naIyZdJJj7gN+KSLnADuAW5ztvwbePuG81wP1IjIP+CHwH4F9ZUoppfwRzCuYhcCzACKyxRiTYYxJ\nFZFGY0wpcEREqgCMMS86++f0dwwwH/iac97ngW8DvwK+BMwEpp/wvA87f18K/C54LzF8HKpv5cX3\n9hAf62VMbjKTSjJJTYx1O1ZEOFjXwkvv7sHrsYiL8ZKU4CMnPYGc9ARGZScRFxOa8x6UCrZgFph8\nYE2f72udbY3O19o+j9UAZUD2AMckiUhbn30LAESkyRjT3/PWOo93G2N6jDGxItI+UNCMjMRhTTLq\n8nhoaG7DFGcO+RzBkJOTcvzvb244wBsf7Dv+vc9rcc70Qq6/aAL5WUmuZgslQ8n16tp9LF9X3e9j\nHo/F2PxUTHEGMyfkctr4HOJjT/3XLlTfLwjdbKGaC0I3W6BzjeRSMSdbUmCgx/rbftKlCYay/3CW\nbcjJSeHHD69m294G7rphBuOL0od8rkDKyUmhtrbp+PdNTccAuOLsscTGeFmxYT/L1uxlxbpqPnNu\nKRfMLsKyTvWtDUy2UDHUXE3N9mefL15kGJOXQuPRdmobWqmpa2X3gSZ2H2yiorqBl96pJMbnYUpJ\nJvOmFTCtLAuvZ/BW6lB9vyB0s4VqLgjdbEPJNVhBCmaBqca+mug1CruDvr/HCp1t7QMc02yMSRCR\n1j77Dva865wOf+tkVy+BsG1vAwC//ftmfvjlM4kJ0aUgAMaNTmNKSRYXnzmG9zYf5LHXtvPX13dQ\nsb+RWz81MaSzh7pR2UmUjkr9xPbOrm527W9k3Y7DfLjjEB9st/+kJcVy3vRRLJpVRHJCjAuJlQqu\nYA5TfgXoHdk1A6gWkSYAEakEUo0xY40xPuAyZ/+BjlmKPSAA5+uSQZ73GufvlwPLAvia+lVWaP+n\ncqjhGK+t2TfI3qHBY1nMnZzPv99yBuNGp7FqSw0PPLme9o4ut6NFHJ/XQ/nodK6eX8YPvnQm99w0\nm4UzRtPR2c1zKyr59v+u4NGl22lobhv8ZEqFkaBdwYjISmPMGmPMSqAbuMMYcxPQICLPALdhDzEG\neExEtgHbTjzGefwe4GFjzFeB3cAfjTFe4DUgHSg0xryBPdrsMeACY8zbQBtwU7Be44kS43y8+O5u\nFs4sDJsrgfTkOL5z3en86tmNfLjjEL94egN3XjPNr6YbNTTF+SkU56dw9fwy3lxXzcur9vDq+1Us\nX1/NpXOKuWh2EbE6MEBFgKD2wYjIXSdsWtfnseXAXD+OQUT2Axf08xTzB3jqm/1PGRhej8V5p4/i\npXf38O7mg5wzbdRIRxiyGJ+H26+awi+e3sD6nYf5y6vb+fyF40esTyZaxcV6uXB2EQtmFLJ8XTV/\ne3sXzyyv4M0P93HDBeM5vTzH7YhKDYt+TA2ghTNG47EsXg+TZrK+fF4PX71iMqNzkln2wT7e/PBk\n3VwqkHxeDwtmjOY/vjKXS+cU03i0nf95agO/enYjDUeD2n2oVFBpgQmgzNR4ppZmsvtgE/sOHXU7\nzilLiPNx59XTSIr38ehr29lX2+x2pKiSGO/j6vll3HPzGZQVprJ6aw3f+793eWfD/sEPVioEaYEJ\nsDmT7UFw72464HKSoclKi+fmSyfS0dnNg89toqNTO/1HWmF2Et+9YSbXLyqno7ObH/1hFX9+ZZv+\nLFTY0QITYNPLs4mL9fLe5oP09Az5TsyumjE+h/nTR7G39igvvLPb7ThRyeOxWDSriH+7cRZj8lN4\nbe1efvDwGmrqW92OppTftMAEWFyMl+njsjnUcIy9teHXTNbrmvPHkZESx4vv7mb/4fB9HeGuMCeZ\nn955LueeNoqqmmZ+8Mf32bq7zu1YSvlFC0wQnFaWBcC6HYdcTjJ0CXE+rl80ns6uHh5eImF7NRYJ\n4mN93HTJBG682NDa1slPH/vwY8v+KBWqtMAEwZTSLCwL1u0M3wIDMGN8NtPHZSNV9ayR2sEPUEF1\n3vRCvn3ddBLifDz8svD46zu08KuQpgUmCJITYigvTKNiXyONLeE7zNSyLD67YBxej8WTb+6ks6vb\n7UhRz4zJ4O4bZ1GQlciSVXv43Qtb6OrWn4sKTVpggmRqWRY9EPbt5fmZicyfXkhNXSvLtFkmJGSn\nJ3DXDTMoKUhlxcYD/OKpDbTpEj8qBGmBCZIJxRlA+BcYgMvnjSUhzsvzKyppbet0O44CUhJj+c7n\npjO5JJN1Ow/z8yfW0dauRUaFFi0wQTI2P4X4WC9b99S7HWXYUhNjuWj2GJpbO7RzOYTEx9oTY2ea\nHLbuqefnT67TKxkVUrTABInX42F8UToHjrRQ1xT+q+QumjWahDgvS1bt0U/KIaR3iZ+Z450i84QW\nGRU6tMAEkRlj33xM9oR/M1lifAyLZhbR1NLBmx/qVUwo8Xk9fPXKycxwiswDT67XWf8qJGiBCaIJ\nY5x+mAhoJgO4YHYRcbFeXnpvj943JsT4vB6+duVkTi/PZsvuOh58bjPd3TqEWblLC0wQjclLJjbG\nQ0V1g9tRAiI5IYYFMwppONrOu5sPuh1HnaC3yEwYk87abbU8/LJOkFXu0gITRF6Ph5L8VPbVHo2Y\n0VeLZhbh9Vi8urpK//MKQTE+L19fPI0xecksX1fNM2/tcjuSimJaYIKstDCVHqByf6PbUQIiIyWO\n2RNz2XfoKJsqj7gdR/UjIc7HNz87ndz0BP6+spLX1ux1O5KKUlpggqxsVBoAO6sjo8AAXDi7CIBX\nVle5nEQNJC0plm9dN53UxBj+snQbGyoOux1JRSEtMEFWOioVgIoIKjBj81MZPzqNjRVHwvLGatEi\nNz2Bry+ehtfj4VfPbtQbyKkRpwUmyNKT48hKjWdndUNE9VlcMHsMAMvWavNLKCsrTOPWT03kWHsX\nP39yfVivjafCjxaYEVA6KpWmlg5qG465HSVgppdnkZ4cyzubDujEyxB35qQ8rpxXwqGGY/ziqQ06\nR0aNGC0wI6CkwG4m232gyeUkgeP1eDhn2iha27pYtVWHLIe6K84ey5mT8tixr4E/v7rd7TgqSmiB\nGQFj8pIB2HMwcgoMwLmnjcIC3vyw2u0oahCWZXHzJROOD19evk5/Zir4tMCMgDF5KQDsjrACk5UW\nz9SyLCqqGyOueEai2Bgvd1w1laR4H396ZRu7ImTovApdWmBGQHJCDFmp8ew50BRRHf0A500fBcCb\n+ok4LOSkJ/CVKybT1dXN/z6zgSbt9FdBpAVmhIzJS6axpYP65sj6hZ5WlkVGShzvbjqg65OFiaml\nWVx5TgmHG9t48LlNumaZChotMCOkON9uJou0piSvx8NZU/Jpbeviwx2H3I6j/HTZWWOZPi6bzZV1\n/H1lpdtxVITSAjNCIrUfBmDu5HwAVm484HIS5S+PZXHrZRPJSo3jbyt2RcQtJVTo0QIzQorzeq9g\nIm829ajsJEoKUthYcYSG5vC/uVq0SIqP4atXTMHC4jfPb9b+GBVwWmBGSHpyLKmJMRHXRNbrrCkF\ndPf08J4u4x9Wxo1O46pzS6hrauN3L2yJuEEoyl1aYEaIZVkU5SZzqOFYxCzd39cZE3PxeixWaDNZ\n2LlkTjGTxmawbudhlr6vS/+owPEF8+TGmPuBOUAPcKeIrO7z2CLgR0AX8KKIfH+gY4wxRcAjgBfY\nD3xBRNqMMTcA3wC6gd+IyEPGmFHA74A4Z/9visiaYL5OfxXmJLOpso7qQ0cpK0xzO05ApSTGMq0s\niw+2H2LPwabjfU4q9Hksiy9fNol7freKx5ftYHxR+vFBKUoNR9CuYIwx5wHlIjIXuBV44IRdHgAW\nA2cDFxpjJp3kmPuAX4rIOcAO4BZjTBJwN7AImA980xiTCXwLeEZEzgfuAn4YrNd4qgpzkgDYG6Gr\n2p41xe7sf2eTXsWEm7TkOL502SS6unv4v79v1vXKVEAEs4lsIfAsgIhsATKMMakAxphS4IiIVIlI\nN/Cis/9Ax8wHnnPO+zx2UTkTWC0iDSLSCqzALlaHgCxn3wzn+5AwOsdeMmZf7cgvcT8SLevTyrJJ\niPOxaksN3VHVlh8Zr3VKaRYLZhRSfegoT71Z4XYcFQGC2USWD/Rtmqp1tjU6X2v7PFYDlAHZAxyT\nJCJtffYtGOAcBcD9wCpjzBeBVGDeYEEzMhLx+bx+v7AT+XxeLAtyck7erJCSmmAHbTg26L6B0Pc5\nkpJiAUhPSwzqc581rYDXVldx+GgHk0qyBtxvJF7/UAwlV2JiHADp6cF7b0fq/brtmulIVT2vrK7i\nvFlFTBuXM+gxkfSzHCmhmi3QuYLaB3MCawiP9bd9sH2/AzwuIj80xlwG/AT4zMmC1dW1nOzhk8rJ\nSaHTmcFeWzv4CLGc9Hgqqxv82nc4cnJSPvYcR4/aQ1AbGlqD+tzTSjJ5bXUVr7xTSU5yrF/ZQsVQ\nc7W02J996utbqK3t/zUPx0i/XzdfMpEfPbKG//7zGu695UwS4wf+byLSfpYjIVSzDSXXYAUpmE1k\n1dhXGb1GYXfQ9/dYobNtoGOajTEJg+zbu/1sYImz7VVg1nBfSCAVZttLxjQejcw5BxOLM0hOiOH9\nrTW6BEmYKh2VymVnFXO4sY1Hl25zO44KY8EsMK8AVwMYY2YA1SLSBCAilUCqMWasMcYHXObsP9Ax\nS7EHBOB8XQK8B8w2xqQbY5KxC8tb2IMAznT2nQ2E1M0vejv6I/X2tT6vhxnjc2g42s62qnq346gh\nuuyssRTnp7Bi4wHWSO3gByjVj6AVGBFZCawxxqzEHg12hzHmJmPMVc4utwGPYheFx0RkW3/HOPve\nA9xojHkLyAT+6HTs3wW8jF2A7hWRBuyhz5cZY94AfgB8M1ivcSiOjySL4HvZnzExF4BVW2tcTqKG\nyuf18OXLJhHj8/Dwy1t1lr8akqD2wYjIXSdsWtfnseXAXD+OQUT2Axf0s/1J4Ml+9r10iJGDbnS2\neyPJRooZk05qYgxrpIYbLijH69H5vOFoVHYSV51TyuPLdvDo0u185YrJbkdSYUZ/80dYflYiXo/F\nvkOR2UQG9grLMyfk0tTSwdbd2kwWzi6cXURJQSrvbj7Ih9tDZsS/ChNaYEaYz+shLzOR6kNHI3rd\npzMm2M1kq7fq2mThzOOxuOXSCXg9Fg+/vJWWYx1uR1JhRAuMCwqyEmlt66IhQkeSAZSPtpvJPth+\nSEeThbnCnGQuP3ss9c3tPL5sh9txVBjRAuOCgqxEAPZHcEe/x2Nx+vgcmlo62L5Xm8nC3aVzihmd\nk8zydfvZVHnE7TgqTGiBcUFBpj2SbP+RoU/wDAczxtuzwNdu07b7cOfzerjlUxPwWBZ/fGkrx9oj\nb0VwFXhaYFyQ33sFcziyC8zE4gwS4rys3VYb0f1N0WJsfioXnzmGQw3HeHq5rlWmBqcFxgX5mXaB\nOXA4cpvIwP7UO60sm8ONxyLyTp7R6Mp5Y8nLTOS1NXupPNDodhwV4rTAuCAhzkdGSlzEN5FB32Yy\nnQ0eCWJ8Xr54kaGnB/74ktDV1e12JBXCtMC4pCArkSONbRHflj21NBOf18Pa7VpgIsXE4gzmTs5n\n98EmXlixy+04KoRpgXFJb0f/gQi/iomP9TGlJJN9tUc5GOGvNZpcu2AcSfE+/rRkC0caj7kdR4Uo\nLTAuiZaOfoDTx2cD2kwWSVKTYrnm/HG0tnXx6NKQWk9WhRAtMC4piKICM31cNpalBSbSzJtWwKSS\nTNZsq+XDHToUXX2SFhiXFGQ5TWQRPpIMICUxFlOUzs7qRhqa2wY/QIUFj2Vx+9Wn4fVY/PkVoa29\ny+1IKsRogXFJenIs8bHeqBhJBnDaOLuZbP3Owy4nUYFU7MyNOdzYxt+0w1+dQAuMSyzLoiArkYNH\nWqJira7eArNOC0zEueysseSkx/PKqir2RuiN9NTQaIFxUX5mEp1dPRxqaHU7StDlZyaSl5nIpl1H\naO/QppRIEhfj5YYLxtPd08NfXt2mqzao47TAuKi3o786Cjr6AU4ry6Kto4uNehUTcaaVZTN9XDZb\n99SzaoveyVTZtMC4KM9ZMqYmSvphpjvNZKs2H3A5iQqG6xaV4/N6eOz17RE/gVj5RwuMi/IyEgA4\nWBf5TWQA40ankRDnY/XmA9qMEoFy0xO4dM4Y6pvbeX5FpdtxVAjQAuOivAxn0csouYLxeT1MLc2k\npq6VfbWRPzw7Gl06p5jstHheWV3F/igYgq9OTguMi+JivaQnx1JTFx0FBvqOJtOJeZEoNsbLdQvL\n6erWDn+lBcZ1eRn2opcdndExsmpqaRYeC535HcFOL89mSkkmmyrrdPWGKKcFxmV5mYn0ADVR0g+T\nnBDDxJIsKvY10tjS7nYcFQSWZXH9BePxeiz++tp22nRYetTSAuOyvEy7o//AkegoMABnTMqjB9ig\nw5UjVn5mIhedYc/wf+Gd3W7HUS7xq8AYYzzGmPxgh4lGvR390dQPM3uS/U9pnTaTRbTLzxpLRkoc\nS97bHVX/vtVHBi0wxpiFwE7gDef7+40xlwU5V9T4aKhy9PwCjs5NJic9no27jtCpd0SMWHGxXq5d\nMI7Orh4eX7bT7TjKBf5cwfwQmAPs7/P994KWKMrkZiRgAQejqInMsiymlWVzrL2L7Xsb3I6jgmj2\nhFzKR6exdlstW3bXuR1HjTB/CkyziBzs/UZEDgHaOxsgMT4vmanxHIiiKxiwR5MBbKjQfphIZlkW\nn1tUjgU8unR7VCzsqj7iT4FpNcacB1jGmAxjzG2A3iM1gPIyE2hobo+q5TUmjEknxufRAhMFxuan\ncvbUAvbWNrN8XbXbcdQI8qfA3A58B5gN7AAuBr4czFDR5qOO/uhpJouN8WLGpLOv9qje0z0KLD6v\nlLhYL08vr6DlWIfbcdQI8fmxT5mIfKxT3xjzaWDQsYfGmPux+296gDtFZHWfxxYBPwK6gBdF5PsD\nHWOMKQIeAbzYfUFfEJE2Y8wNwDeAbuA3IvKQc45vA58HOoDb+z5vKOpd9PJgXStj8lJcTjNyppZm\nsbHiCBsqDnPe9EK346ggSkuO47K5xTz1ZgXPrajkuoXlbkdSI2DAKxhjzFhnBNn9xpjzjTELnD8X\nAT8b7MROs1q5iMwFbgUeOGGXB4DFwNnAhcaYSSc55j7glyJyDvZV1C3GmCTgbmARMB/4pjEm0xgz\nGbgOmAV8FQj5EW+9I8miZU2yXtOO98MccTmJGgkXzi4iOy2e19bs1XXKosTJmsgKgGuBsdj/kf+b\n8+efgV/7ce6FwLMAIrIFyDDGpAIYY0qBIyJSJSLdwIvO/gMdMx94zjnv89hF5UxgtYg0iEgrsAK7\nWF0GPC4inSKyVkTu8SPrsAy32zLalu3vlZeZSG56ApsrdbhyNIjxebl2gb1O2WOv73A7jhoBAzaR\nicg7wDvGmBdF5Nm+jxljzvLj3PnAmj7f1zrbGp2vfRcpqgHKgOwBjkkSkbY++xYMcI4C7ILYZYxZ\nAsQA3xKRdScLmpGRiM/n9eMl9S/G5wUscnKG1ryVkZmEx2NxpLl9yOfoT99zJSXGApCWlhDQ5xiq\n3gxnTM7n7yt2cai5g6nOQphuGsp7k5Bgv7fp6YlBe29D4Wc2kFPJdlF2Mm9t2M/6HYeoOtzKjAm5\nIZFrpIVqtkDn8qcP5nVjzO3Y//kDxAE3A6NO8bmsITzW3/bB9rWw+2ouwb6i+S32AIUB1Q1jiHBO\nTsrxhSpra5uGfJ7s1Hj21jQN6xwn5up7rqNH7frc0NgasOcYqr7Zxo2y/0G/tbaK/LQ4N2N94j3z\nV4uzplpDQwu1tbGBjjXkXCNhKNkWn1vKhp2H+PXT67j3ljPweQO/YlWkvWcjYSi5BitI/vxkHwOm\nYReVFOwmqNv8OK4a+yqj1yg+mqx54mOFzraBjmk2xiQMsm/v9oPAchHpEZG3sa9oQl5eZiJNLR1R\nN8LGjMnA59XhytGkKDeZ804bxf7DLbzxwT6346gg8qfAxIvI14DdIvId4Hzgs34c9wpwNYAxZgZQ\nLSJNACJSCaQ6Awl82EXrlZMcsxR7QADO1yXAe8BsY0y6MSYZ+2rlLeAl4CLnHBOAKj+yui7a7m7Z\nKy7Gy4Qx6ezV4cpR5dPnlpIQ5+Nvb++iuTW6PlRFE38KTJwzYstjjMkSkSPY/SUnJSIrgTXGmJXY\no8HuMMbcZIy5ytnlNuBR7KLwmIhs6+8YZ997gBuNMW8BmcAfnY79u4CXsQvQvU6H/7vAbmPMO8Dv\n+5wjpB0fqhxlHf3w0az+jbt0NFm0SE2M5Yqzx3L0WCd/e2uX23FUkPjTB/Mw9sTK3wJbjDG1wHZ/\nTi4id52waV2fx5YDc/04BhHZD1zQz/YngSf72X4PdlEKG9F6BQMwtSyLR1/bzoadhzn3tFPt2lPh\nauHM0bzxwT6WfbCPBTMLKchKcjuSCjB/rmAeFJGficjDwOnYExivGuQYdYpyj0+2jL4rmLyMBHLS\n49mkw5Wjis/r4bPnj6O7p4fHddhyRPKnwLze+xcR2SciH4iIrlgXYFmpcXg9VlStqtzLsiymldqr\nK+/Q1ZWjyvTybCaMSWfdzsNsqtQm0kjjT4H50BhznzHm4j6z+RcEPVmU8Xo85KQnRO2NmaaWZQK6\nunK0sSyLaxfYqy0/9toOXW05wvhTYKYD5wD/wkez+fV+MEGQm5HA0WOdUTmqRocrR6/i/BTOmprP\n3tpmVmzYP/gBKmwM2skvIuePRBDVu6ryYQ7WtZCckOZ2nBHVO1x5464jHGk8RmZqvNuR1Aj6zLll\nrN5aw9PLK5g9MZf4WH/GH6lQF/gptGrI8jLtkWQ1UdgPAzpcOZplpMRx8RljaDjazkvv7nE7jgoQ\nLTAhpPe+MNE4kgzs4cqg/TDR6pIzi0lLjuXlVXt00m2E0AITQnKduTDRdOOxvvIyEshOi9fVlaNU\nXKyXxeeW0d7ZzVNvVrgdRwXAoA2dzuz5E4d2dAIC/EBEdDGhAMlKjcfntaL2CsayLKaWZbFs7T52\n7mvAjMlwO5IaYWdNzWfpmire2XSARbNGU1KQ6nYkNQz+XMEsxV7P62fAT4EK4G1gG/ZSLCpAPB6L\nnPQEDh5ppacnOodraj9MdPM4w5YBHnt9R9T+HkQKfwrMPBG5QUSeFpG/ichNwEwRuR8I/NrkUS4v\nI5GWtugcqgwwcUwGPq/Fhp3aDxOtJhZnMH1cNtuq6lm77ZDbcdQw+FNgco0xx+8EZYxJA4qNMelA\ndI2lHQHR3g8TF+tlfFE6e2qaqW9uG/wAFZGuOb8Mr8fiiWU7tD8ujPlTYH4ObDXGvG+MWY3dRPZ7\n7CX2HwxmuGj00aKX0dkPAx81k+losuhVkJXE+acXUlPfyutr9rodRw3RoAVGRH4HlABfBW4HykXk\npyLyJxH5dbADRpvji15G6VwY6NMPU6H9MNHsinklJMb5eG5FZdQ2GYe7QQuMMSYfuAW4Argc+IYx\n5r5gB4tWegUDBVmJZKXGsWnXEbq6tXkkWiUnxHDF2WNpaevkubf1njHhyJ8msheA04BuoKvPHxUE\nmanx+LyeqO2DAWe4cmkWLW2dVFQ3uh1HuWjBzNHkZiSw7IN9HIjCm/GFO38W/GkWkVuCnkQB9jDN\nnPR4DtbZQ5Uty3I7kiumlmbxxofVbKg4QvnodLfjKJf4vB6umT+OXz6zgSeW7eDri6e5HUmdAn+u\nYN517m2vRkheRiKtbZ00RXG784TiDLweSzv6FTPGZzN+dBofbD/Elt11bsdRp8CfAnMxsMEYU22M\n2WOMqTLG6Gp0QRTti14CJMT5KB+dxu4DTTQcbXc7jnKRZVlcu9CZfPnadr1nTBjxp8BcAYwD5mLf\nF2ae81UFSbQvetmrd/HLjXoVE/VKClKZOzmfPTXNrNx4wO04yk8DFhhjzCXOXxcO8EcFSe7xkWTR\newUDumyM+rjF55US6/Pw1PKdtLXrOKNwcLIrmN7etHMG+KOCpPcKJlpvn9yrMDuJjJQ4NlYc1mYR\nRWZqPBedMYaG5nZeem+323GUHwYcRSYi/+V8vXnk4iiAjNQ4YnyeqJ5sCb3DlTNZvm4/u/Y3Ulao\nKxNFu0vmjGH5umqWrNrDedMLyUiJczuSOgl/Jlp+zhjzgTFmt9PJv0c7+YPLY1nkpidwsK4l6leT\n1WVjVF/xsT6uOreU9o5unl6+0+04ahD+dPLfC9wJnIs2kY2Y3IwEjrV30dQSvUOVASYWZzrDlbUf\nRtnmTS1gdE4yKzccYPeBJrfjqJPwZ6LldhFZHvQk6mP6jiRLTYreuyIkxvsoK0xje1U9jS3tpCZG\n73uhbB6PxXULx/GTv37Io69t51+uPz1qJySHOn8KzEpjzI+AN7DvZAmAiLwerFAKcp25MAePtEb9\nTPappZlsq6pn064jzJ2c73YcFQImjc1k+rhsPtxxiDVSy6wJuW5HUv3wp4lsEfYcmO8C/+b8+V4w\nQymdC9PXR6sraz+M+si1C8ZTRoBbAAAgAElEQVTh9Vg8vmwHHZ06bDkU+XMF808isjboSdTH5EX5\njcf6KspNJi05lo27jtDd04NHm0MUkJeZyKJZo3l5VRWvrK7iU3PHuh1JncCfK5ifBD2F+oT0FGeo\nsl7B2MOVS7JoaunQTl31MZefNZbkhBj+/s5uvQNqCPLnCmaPMeYN4F3g+KJQInJ3sEIpZ6hyRkLU\nr6rca2pZFm9v2M+GisOUFKS6HUeFiMT4GD5zbikPvyw8vbyCWy6d6HYk1Yc/BWaX86cvvyZnGGPu\nB+Y4+98pIqv7PLYI+BH2vWVeFJHvD3SMMaYIeATwAvuBL4hImzHmBuAb2Peq+Y2IPNTn/HnAVuAq\nEXnDn7yhJi8jkX21R2k82k5acnRPKJs8NgOPZa+ufMXZJW7HUSHknNMKeH3tXlas38/CGaMpzk9x\nO5Jy+HPL5HtP/AMM+hM0xpyHfXvlucCtwAMn7PIAsBg4G7jQGDPpJMfcB/xSRM4BdgC3GGOSgLux\nByHMB75pjMnsc/4fAxWD5QyEYM2FzNM1yY5LjI+htDCViupGvX2u+hivx8PnFpbTAzy6dFvUT04O\nJf7M5L/AGLPaGFPh/NmHvYT/YBYCzwKIyBYgwxiT6pyzFDgiIlUi0g28yEeLaPZ3zHzgOee8z2MX\nlTOB1SLSICKtwArsYoUxZgHQBGzwI2fIyg3g7ZMj4VduamkWPT2wuVInXaqPmzg2k9PLs9m2t4E1\nUut2HOXwp4nsB8DXgZ9hX1VcC7zlx3H5wJo+39c62xqdr33/FdQAZUD2AMckiUhbn30LBjhHgTEm\nFrgHuNLJPKiMjER8Pq8/u/YrJsaDx4KcnMBempuSbEBobusa0rn7HpOUZDexZaQlBjznUAwlw7kz\ni3hmeQXb9jXyqXPHBSHV0HIlOpM/M9KTgvbehsLPbCChku1ri0/jjh+/zpPLK1g4Z2zI5OpPqGYL\ndC5/CkyjiLxrjGkXkU3A3caYl4BXT/G5TtZLPdBj/W0fbN+7gP8TkXpjjF/B6oZxhZCTk0JHRzc9\nQG1tYEc4xTnXl7v21p/yuXNyUj52zNGjdn2ub2gJeM5TdWI2f6XEekhNjOH9LQc5WNMY8OHKQ83V\n0mKPfamvb6G2NiagmWDouUZCKGWLARbNLGLJqj38bflO5k8rcDtSv0LpPetrKLkGK0j+DFOOMcbM\nA+qMMTcaY2YD/vSyVmNfZfQahd1B399jhc62gY5pNsYkDLJv7/aLgH8wxrwLfAr4X2PMZD/yhpz0\n5FhiYzw6F8bhsSymlGbReLSdqoPNbsdRIeiys8aSkhjDE69t02HLIcCfAvNV7NFb3wFuAP4Pe/TX\nYF4BrgYwxswAqkWkCUBEKoFUY8xYY4wPuMzZf6BjlmIPCMD5ugR4D5htjEk3xiRj97+8JSJni8gc\nEZkDvADc7lx5hR3LsshNTzw+VFnp6srq5BLjfVx1TimtbV08vXxExviok/BnFJlg97k0isiFIjJd\nRB7247iVwBpjzErs0WB3GGNuMsZc5exyG/Coc+7HRGRbf8c4+94D3GiMeQvIBP7odOzfBbyMXYDu\nFZEG/196eMjLSKCto0vvS++YXJKJZWmBUQM797RRjC1IZcX6/Tox12WD9sEYYxYCvwXagAnOPJXX\nROTvgx0rInedsGldn8eWY69xNtgxiMh+4IJ+tj8JPHmS579psIyh7qNFL1tIj/K5MADJCTGUFqSy\nc18jLcc6SIwPfJ+HCm8ej8WXrpzC9369kkeXbuNfbpgR9ROV3eJPE9kPsSc+7u/zvS52OUI+WvRS\n+2F6TS3Norunh82VdW5HUSHqtPKc48OW39dhy67xp8A0i8jB3m9E5BB9loxRwaWLXn7S1DK7H2a9\nNpOpk/jsgnH4vBaPvb6dtg5dbdkN/hSYVmeGvWWMyTDG3AYcC3Iu5cjVZfs/oTg/heSEGDZWHNbB\nD2pAeRmJXHTGGI40tvHCO7vdjhOV/Ckwt2OPIJsN7MSexf+VYIZSH0lPjiUuxsvBI3oF08serpxJ\nfXM7VTU6XFkN7LK5Y8lIiWPJe7v1Q5oLBu3kF5Eq7GHEygXW8VWVW3RV5T6mlmbx7qaDbNx1hDF5\noTkrWrkvLtbLtQvG8eu/beKvS7dz5zWnuR0pqgxYYJwhwQO2P4jIuUFJpD4hLyOBqppm6pvbyUjR\nkWTgDFcG1u84xKVzit2Oo0LY7Am5vPHBPtbtPMyHOw4xfVy225GixsmuYHSkWIjIy7T7YWrqWrTA\nOFITYyktTGX7vgaaWztITtDhyqp/lmVxwwXjued3q/nr0u1MHptBzDDWHlT+G7DAiMibIxlEDSw3\n/aNl+82YDJfThI7p47LZua+R9TsPcdaU0Fx3SoWGwpxkFs0azSurq1iyqorLzxrrdqSo4E8nv3JZ\n7xXMwSPaSdnX9PIcAD7cfsjlJCocXHF2CalJsbywspLDDToQdiRogQkDOhemf6OyEslNT2DjriN0\ndHa7HUeFuMR4H9fML6O9s5vHXt/udpyooAUmDKQmxRIX69VhliewLIvp5dkca+9CqnRWvxrc3Cn5\njCtM432pZZPeuC7otMCEAcuyyEtPoKaulW6dWPgxvSOCtJlM+cPjdPhbwF9e3UZnl175BpMWmDCR\nm5lIe2c39U16j4u+xo1OIzHOx4c7DumsfuWX4vwU5p9eyP7DLby8ao/bcSKaFpgw0dsPo4tefpzP\n62FaWRZHGtt0Vr/y22fOKyU1MYbnVlRSU6+/U8GiBSZM9K6qXKP9MJ8wvdxpJtuhzWTKP0nxMVy3\nsJyOzm7+9Iro1W+QaIEJE7l6BTOgKSVZeD2W9sOoU3LmpDwmjc1gY8URVm+tcTtORNICEybynbkw\nBw7rFcyJEuN9mDHpVB5ook77qJSfLMviCxcZfF4Pjy7dTsuxTrcjRRwtMGEiJTGGpHgf+w8fdTtK\nSDrNGU22TpvJ1CnIy0jk8rOKaTjazlPLd7odJ+JogQkTlmUxKjuJmvpWnVTYj+PDlbXAqFN08ZnF\nFGQl8sbafVRUN7odJ6JogQkjBVlJ9PTokjH9yUlPYHROMpsrj9Dapk0dyn8xPg9fvMjQAzy8ZCtd\n3foBLlC0wISRUdlJAFRrM1m/Zk3IobOrR5vJ1CkzYzKYN7WAPTXNLH1/r9txIoYWmDAyKtvu6K8+\npAWmPzNNLgDvS63LSVQ4uub8MpITYnjmrQqdGxMgWmDCyKgs+wpmv44k61dhdhIFWYlsrDhMW3uX\n23FUmElJjOX6ReW0d3Tzx5e26tyYANACE0YyUuKIi/VqE9lJzDS5tHd2s6HisNtRVBg6c1Ie08dl\ns2V3HcvXVbsdJ+xpgQkjlmUxKiuRg0datCNyALOMfY+Y90UnzqlT1zs3JiHOy2Ov7+BIo943Zji0\nwISZUVlJdHb1UFuv//D7U5SbTE56POt2HqajU5vJ1KnLSInj2gXlHGvv4uGXdRmZ4dACE2aOjyTT\njv5+WZbFTJNLW3sXG3fp/T7U0JwzrYBJYzNYv/Mw72466HacsKUFJswUHO/o1wIzkFm9o8m26mgy\nNTSWZXHTxROIi/Hyl6XbaGjWJYiGQgtMmNGhyoMrKUghMzWOD3cc0htKqSHLTk/g6vllHD3WyZ9e\n2aZNZUOgBSbMZKcl4PN6qNahygOyLIuZ43NpbetkkzaTqWE4f0Yh44vSWbOtlnc2HXA7TtjxBfPk\nxpj7gTlAD3CniKzu89gi4EdAF/CiiHx/oGOMMUXAI4AX2A98QUTajDE3AN8AuoHfiMhDxhgf8BBQ\n5ry+b4vI28F8nSPJ47EoyEpk/+GjdPf04LEstyOFpDMm5fLq+1W8t+Xg8YUwlTpVHsvi1k9N5O7f\nreLPr27DFGWQlRbvdqywEbQrGGPMeUC5iMwFbgUeOGGXB4DFwNnAhcaYSSc55j7glyJyDrADuMUY\nkwTcDSwC5gPfNMZkAl8AjorIPOcc/x2s1/iRkb10LsxOor2jm0MNOpJsIKUFqeSkx/PBtkM66VIN\nS056AtcvKqe1rYuHXthMtzaV+S2YTWQLgWcBRGQLkGGMSQUwxpQCR0SkSkS6gRed/Qc6Zj7wnHPe\n57GLypnAahFpEJFWYAV2sfoT8C1n31ogK4iv0RWjc5MB2Ku3CB6QZVmcOSmfto4uXWFZDdu8qQWc\nXp7N1j31vLq6yu04YSOYTWT5wJo+39c62xqdr32H+NRgN2llD3BMkoi09dm3YIBzFIhIB9DhbPsG\n8JfBgmZkJOLzef17Vf3w+bxYlkVOTsqQz3EqJo/L4ck3dnLkaMdJn7PvY4mJcQCkpyWOWM6TGYkM\nl84r5e8rK/lgx2EuO2+cX8cMJVdCQiwA6enBe29D4Wc2kFDNFuhc//T5WfzDj5fx9PIKzplRRHFB\n6pDPFS3vWVD7YE5wss6CgR7rb7tf+xpj7gBmAJcPFqxuGPe5z8lJobOzC3p6qK1tGvJ5TkVqnF0M\nZdfhAZ8zJyflY4+1HLXrc0NDy4jlHMiJ2YIl3gNj8pJZs/Ugu/YcITkhJii5WlraAahvaKG29uTP\nMRQj9X4NRahmC1auL15keOCp9fzXw6v53hdnEeM79UagSHrPBitIwWwiq8a+yug1CruDvr/HCp1t\nAx3TbIxJGGTf3u0YY27FLiyfdq5oIkp6cizJCTFU1WoT2WDmTMqnq7tHl45RATG9PJtzTyugqqaZ\nZ5ZXuB0n5AWzwLwCXA1gjJkBVItIE4CIVAKpxpixzqivy5z9BzpmKfaAAJyvS4D3gNnGmHRjTDJ2\n/8tbTv/O14DPiEhE9oJblkVRbjK1da0ca9eba53MGRNzsYD3dDa2CpDrFpaTl5HAklV7WL9T+/dO\nJmgFRkRWAmuMMSuxR4PdYYy5yRhzlbPLbcCjwFvAYyKyrb9jnH3vAW40xrwFZAJ/dDr27wJexi5A\n94pIA/Al7I79F40xbzh/YoP1Ot0yOieZHmBfrU64PJnM1HjGF6UjVfW6cKEKiPhYH7d9ego+r4ff\n/n0LdU06y38gQe2DEZG7Tti0rs9jy4G5fhyDiOwHLuhn+5PAkyds+1fgX4cYOWwUOSPJqmqaKStM\nczlNaJs7JR+pqmfFxgNcftZYt+OoCDAmL4VrF4zjz69u48HnNvGdz03H69F56yfSdyRMHS8w2g8z\nqNkTcomN8fD2+mqdw6ACZsGMQmaOz2FbVT3Pr6h0O05I0gITpkZlJ+KxLKp0LsygEuJ8zDa51NYf\nY9ueerfjqAhhWRY3XzqBrNR4nl9RyZZKXZboRFpgwlSMz0tBViJ7a5r1U7kf5k0rAOCt9fsH2VMp\n/yXGx/C1Kyfj8Vg8+Nwm7ec7gRaYMFaUm8yx9i4O1be6HSXkjS9KJzcjgTVSQ8sxHXmnAqesMI3P\nnj+OxpYO/vfZjXR06grevbTAhLHifHuS0679oTdpK9RYlsW8qQW0d3azaosOWVaBtWjWaOZMzqOi\nupE/v7rN7TghQwtMGCtxlqrYtb/R5STh4eypBVgWvLW+2u0oKsJYlsWNF09gTG4yy9dV8+aH+9yO\nFBK0wISx4rwULAsqtcD4JSMljqmlWeza30TlAX3PVGDFxXi54zNTSYr38edXt7GzusHtSK7TAhPG\n4mK9jMpOYvfBZrq7taPfHwtnjgbgtTV7XU6iIlFOegJfu3IKXd09/OKpDVHf6a8FJsyV5KfS1tFF\n9WGd0e+PySWZ5GYk8N7mGpqcRSqVCqTJJZlce/44Go6287Mn1tPaFr2DSrTAhLmSgt6Ofm3y8YfH\nslgwYzSdXd06ZFkFzQWzizj/9EL21jbz679toqs7OkeWaYEJc2Odjv5KHUnmt3lT84mN8bBs7V5t\nWlRBYVkW119QzpTSTDZUHOYvS7fTE4Xz1bTAhLmi3GR8XkuvYE5BYnwMZ00p4HBjGx9s19VwVXB4\nPR5uu3IKo3OSWLZ2Hy+vir47YWqBCXM+r4fivBSqapr13vOnoLezf8l7u6Pyk6UaGQlxPr5xzWmk\nJ8fy+LIdvLUuuobIa4GJAOOL0unq7tFhkaegMDuJ6eOy2VndyLYqXZ9MBU9majz/dO10kuJ9/GHJ\nVlZG0TwsLTARoLwoHUD/ozxFn5pbDMAL7+x2OYmKdIU5yXzzs9OJjfHy4z+tYVOULIypBSYClI9O\nw0ILzKkqK0zDFKWzcdcRdh/QQRIquEpHpfKPn5kKwC+e2oDsqXM5UfBpgYkASfExFOYks7O6kc6u\n6BwOOVQfXcVUuppDRYeJYzP5ly/OorOrm/ufWMfW3ZFdZLTARAhTlE5HZzeV+kn8lEwuyaSkIIX3\npZYde/UKUAXfnCkF3H7VFLq6evjZE+siurlMC0yEGD/G7ofZEuGfiALNsiwWn1cGwCMvbnE5jYoW\np5fn8PXFU+nugQeeXM/6nYfdjhQUWmAixMTiDCwLNlVE5j/UYJo0NpOJxRmslZqIb7JQoWNaWTb/\nuNjuk/mfp9azYkPkrSyhBSZCJCfEUFqQyo59jbQc63A7Tti5er59FfPEGzv1DqFqxEwpzeKfrp1O\nfKyXh17Ywt9XVkbUvCwtMBFkamkW3T09bK7UT+GnqqQglXOmF7JrfyMrdI0yNYLGF6Xz3c/PJCs1\njqeXV/DIyxIxg3W0wESQqWVZAKzXZrIhueXyycTFeHnijZ00t+pVoBo5o7KT+NcvzKIoN5k3Pqzm\nJ49+QMPR8F/tWwtMBCnOTyElMYb1Ow/TpYs4nrLs9ASumDeW5tYOnn5zp9txVJTJSInju5+fwawJ\nuWzb28B9f1gd9qtzaIGJIB7LYqbJpfFoOxt36CKOQ3HBrCIKs5N448NqNu2K3OGjKjTFx/q47crJ\nXDO/jPrmNv7zT2tZ8t6esO0X1AITYc6cmAvAcr0n+JD4vB6+dNkkvB6Lh17YrE1lasRZlsUlc4r5\n1menk5QQw+PLdvCTRz8Iy7tjaoGJMOWj00lPjmXl+mo6OiOjo3CkFeencOW8Euqb2/n9i1vC9tOj\nCm+TSzK579YzmD4um6176rn7oVW88eG+sPr3qAUmwng8FnMm5dPc2sEaqXE7Tti6dE4xE8ak88H2\nQzz39i6346golZoYy9cXT+XGiw3dPT08vET4j0fWsOdgeKzYoQUmAs2fUYhlwdI1e92OErY8Hovb\nPj2F7LR4nltRyTubDrgdSUUpy7I4b3ohP/zyHM6YmMvO6kbu/cNqHnphM4caWt2Od1JaYCJQbnoC\nZ0zKp0LvdTIsKYmx/OPiaSTE+fjt3zezastBtyOpKJaREsfXrpzCtz57GqOyklix4QDfffBd/vSK\nUFPX4na8fmmBiVBXLywH4LHXd0TUzOCRNjo3+fhM6wef28SS9/bo+6lcNaU0i3tvOYMvXz6JzNQ4\nXl+7j+8++C7/89R6Nu06QncITVHwBfPkxpj7gTlAD3CniKzu89gi4EdAF/CiiHx/oGOMMUXAI4AX\n2A98QUTajDE3AN8AuoHfiMhDxpgY4A9AsXPum0WkIhiv75fPbGDrnnpajnUQ4w2tWj2hOJNZE3J5\nf2sNr2lT2bCUjkrl29edzgNPrefxZTvYsa+BGy4YT0ZKnNvRVJTyeCzmTs5n9oRc3pcaXl1dxQfb\nD/HB9kOkJsVyxoRcppdnUz46jRif17WcQSswxpjzgHIRmWuMmQj8DpjbZ5cHgIuAfcCbxpingJwB\njrkP+KWIPGGM+RFwizHmYeBu4AygHVhtjHkGuByoF5EbjDEXAv8BXBuM15iZEk9OegKdnXFMKE4P\nxlMMy+cWliN76njs9R1kpup/hsNRUpDK3TfO5sG/bWTttlo27jrM2VMLOGNCLu0dXW7HU1HK5/Uw\nZ1I+Z07MY2d1Iys3HuD9rTUsXbOXpWv2EuPzMK4wjTF5yYzOSSY/K5H0pDhSk2KJ8QX/Q3Ewr2AW\nAs8CiMgWY0yGMSZVRBqNMaXAERGpAjDGvOjsn9PfMcB84GvOeZ8Hvg0IsFpEGpxzrADOds7zsLPv\nUuwiFRSfW1ROTk4KtbWhOaIjIyWOO66ays+fXEdtffiNoQ81GSlx/PMNM3h7/X6eX7GLZWv3sWyt\nzjdS7rMsi3GFaYwrTOP6ReVs2V3Hpl1H2FxZx5bddf3exsPrschOT+Dfb5pNXGxwrnKCWWDygTV9\nvq91tjU6X2v7PFYDlAHZAxyTJCJtffYtGOAcH9suIt3GmB5jTKyIDLiwT0ZGIr5hXEbm5KQM+dhg\nyslJIScnhZKiDJ58fTsNze3MmjqKxPgYt6OF9Hs2mMWLUvn0+eWskRo+kBqqa4/i9VqcPjGf+Ljg\n/EqF6vsFoZstVHNB8LMV5Kex4MyxADS3drB7fyOV+xs5cPgo9U1t1De1cay9k4zUePLzU/E5TfyB\nzhXUPpgTWEN4rL/tp7LvYM8LQN0wRmCE6hVM31wxwOcWjAPgaNMxjja5ezUTDu+ZP0pykijJKTn+\nfVNjK8F4VaH6fkHoZgvVXOBOttyUWHJTsrE/w39c3ZGjQ841WEEKZiNcNfbVRK9R2B30/T1W6Gwb\n6JhmY0zCIPt+YrvT4W+d7OpFKaVUcASzwLwCXA1gjJkBVItIE4CIVAKpxpixxhgfcJmz/0DHLAUW\nO+ddDCwB3gNmG2PSjTHJ2P0vbznnuMbZ93JgWRBfo1JKqQEErYlMRFYaY9YYY1ZiDyO+wxhzE9Ag\nIs8AtwGPOrs/JiLbgG0nHuM8fg/wsDHmq8Bu4I8i0mGMuQt4GXtI870i0mCMeQy4wBjzNtAG3BSs\n16iUUmpglk4ag9rapiG/CaHa1huquSB0s2muUxeq2UI1F4RutiH2wZy0jzu0ZgcqpZSKGFpglFJK\nBYUWGKWUUkGhBUYppVRQaCe/UkqpoNArGKWUUkGhBUYppVRQaIFRSikVFFpglFJKBYUWGKWUUkGh\nBUYppVRQaIFRSikVFCN5w7GIYoy5H5iDvZLznSKyegSf+/8DzsH++f0HsBp4BPBi3z/nCyLSZoy5\nAfgG9srUvxGRh5x75PwBKAa6gJtFpCKA2RKAjcD3gddCIZfzfP8MdAJ3A+tDJFcy9u29M4A44F7g\nAPAr7H9X60XkNmff72DfhqJ35fAXjTFpwF+ANKAZuF5Ejgwz0xTgb8D9IvILY0wRw3yvjDGn9fea\nApDr99j30+sAPi8iB9zO1Wf7RcASEbGc70c0V3/ZnOf7IzAOaAKuFpG6YGbTK5ghMMacB5SLyFzg\nVuCBEXzu84EpznNfDPwMuA/4pYicA+wAbjHGJGH/Z7oImA980xiTCVwP1IvIPOCH2AUqkL4H9P4n\n53ouY0wW9u0e5mHfd+jKUMjluAkQETkf+z5IP8f+ed4pImcDacaYS4wxJcB1fV7DfxtjvNj/Kbzh\nZHsa+JfhhHHeg//B/mDQKxDv1SdeUwBy/QD7P8PzgGeAb4VILowx8cB3cW6wONK5TpLty0CtiJwB\nPAacE+xsWmCGZiHwLICIbAEyjDGpI/Tcy/nohmr1QBL2P4znnG3PY/9jORNYLSINItIKrMC+KdtC\n7F9IsG/kdnagghljJgCTgBecTaGQaxGwVESaRGS/iHwlRHIBHAKynL9nYBfmkj5Xw73ZzgdeEpF2\nEanFvifSpBOy9e47HG3Apdh3he01n2G8V8aY2AFe03Bz3Q485fy9Fvt9DIVcAP8K/BLovZPuSOca\nKNvlwJ8BROQ3IvJcsLNpgRmafOx/1L1q+fjtm4NGRLpE5Kjz7a3Ai0CSiLQ522qAgn4yfmK7iHQD\nPc4/nED4KfCtPt+HQq6xQKIx5jljzFvGmIUhkgsR+SswxhizA/uDw7eBupNlOMn23m3DydPp/CfT\n17DeK2dbf69pWLlE5KiIdDlXcndgNxW6nssYMx44TUSe6LN5RHMNlA37d+ESY8wbxpi/OlcqQc2m\nBSYwTnrTnWAwxlyJXWD+wc8sp7r9VPN8EXhHRHYF6PkD9Z5a2J9uP4PdJPX7E87tVi6MMZ8H9ojI\nOGAB8KdhZBiJf4OBeK8C+f55sfuHXheR1/rZxY1c9/PxD1n+Pv9A2wP5c7Wwm2TnY/eTfvcUnm9I\n2bTADE01H79iGYXT3joSnA7E/wdcIiINQLPTuQ5Q6OQ7MeMntjsdeZaItDN8nwKuNMa8C3wJ+LcQ\nyXUQWOl8otuJ3bnZFAK5wG6KeBlARNYBCUD2yTKcZHvvtkAb1s8Q+/ciq599A+H3wHYRudf53tVc\nxphCYALwZ+f3oMAY86bbufo4CLzp/P1lYHKws2mBGZpXsDtlMcbMAKpFZETugeqMHPoxcFmfEUNL\ngcXO3xcDS4D3gNnGmHRntNLZwFtO9t4+nMuBZYHIJSLXishsEZkD/BZ7FJnruZzzLjDGeJwO/+QQ\nyQV2p/mZAMaYYuzit8UYM895/DNOtteBTxljYo0xo7B/sTefkK33dQTasN4rEekAtvbzmobFGfnU\nLiL39Nnsai4R2SciZSIyx/k92O8MQnD9/XK8hD0wCGAmIMHOpsv1D5Ex5j+Bc7GH9t3hfAIdief9\nCvDvwLY+m2/E/k89HrsD+GYR6TDGXA18B7v99H9E5M9Os8JvgXLsjsCbRKQqwBn/HajE/pT0sNu5\njDFfxW5OBHv00eoQyZUM/A7Iwx5y/m/Yw5QfxP7w956IfMvZ9+vADU6274nIa87xf8L+VFmPPVS3\nYRh5ZmL3o43FHvq7z3nOPzCM98oYM6m/1zTMXLnAMaDR2W2ziNweArk+0/vBzxhTKSJjnb+PWK6T\nZLsee6RiAfaw9htF5GAws2mBUUopFRTaRKaUUiootMAopZQKCi0wSimlgkILjFJKqaDQAqOUUioo\ndDVlpQLAGDMdeyj0r4B4EVkbgHOOAiaIyOvGmJsAr4g8NNzzKjVStMAoFQAi8iHwdWPM/8OeMT3s\nAoO9yOVE7KVQ/hCA8yk1onQejFIBYIyZj700ei3QgH1vl5eAXwM52Pds+amI/MWZiFqCfa+Nf8Je\nIua/sCe0JWKvFFyHvShJYmcAAAH2SURBVGqAhT05LhXwicj3jDGfwl5ivcX58xUR2WeMqXT2vcQ5\n/9ecCZl3Ap/vs//nReRwEN8OpQDtg1EqkN7BXjrjxyLyF+xVA5aIyALsVR/uM8bkOPuWAOeLyBrs\n9cduc/b7OfCvzqKhfwAeEZH/7n0CY0wi9gzrxWLfR+Yl53l6tYrIhc62f3S23Ye9tNB52PfzGBX4\nl67UJ2kTmVLBcz72Ok83Ot93YBcWgHdFpLf54ADwE+dGVWl8fEn0E40HDorIXuf7N4Cv9Xn8Defr\nbiDT+ftDwBJjzJPAEyLSd5khpYJGr2CUCp424HYRme/8mSgiq5zH+q7I/AjwnyJyLvYq2SdzYpu2\ndcK2zhMew1kv6tPYNzR7dih3SFRqKLTAKBVY3dj3iQd4G/gsgDEmwRjzv8aY/loN8oBNzgKD1wBx\n/Zyr1zYg1xgzxvl+EfDuQGGMMRlOn0+ViPwK+06LZ5zyq1JqCLSJTKnAeh27ucvCXvX6t8aYt7GL\nxm9EpNMYc+Ix/+Uctxv7VgyPGGO+gb1s+mPGmHagC0BEWo0xtzrb27BXxb31xBP2EpE6Y0wKsNoY\nU4fdTDfg/koFko4iU0opFRTaRKaUUiootMAopZQKCi0wSimlgkILjFJKqaDQAqOUUiootMAopZQK\nCi0wSimlguL/Bw5xvFbIamtQAAAAAElFTkSuQmCC\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7fc144442c50>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"metadata": {
"id": "7inCzCv0cPS-",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "60dce6ce-56d1-457a-f5e6-e66de1f16521"
},
"cell_type": "code",
"source": [
"# Просчитаем метрику точности предсказаний\n",
"x_t, y_t = m.predict_with_targs()\n",
"metrics(x_t,y_t)"
],
"execution_count": 288,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0.93297744"
]
},
"metadata": {
"tags": []
},
"execution_count": 288
}
]
},
{
"metadata": {
"id": "E8uYBfOHPodK",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "09cc471a-f56b-462c-a677-151172305454"
},
"cell_type": "code",
"source": [
"# Получить предсказания тестового набора\n",
"pred_test=m.predict(True)\n",
"pred_test = np.exp(pred_test)-1\n",
"'Ok'"
],
"execution_count": 289,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Ok'"
]
},
"metadata": {
"tags": []
},
"execution_count": 289
}
]
},
{
"metadata": {
"id": "VBMYWe9VRm31",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "1383336d-40a7-4a9b-d42f-33690bde58d6"
},
"cell_type": "code",
"source": [
"len(pred_test), len(y_test)"
],
"execution_count": 290,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(99, 99)"
]
},
"metadata": {
"tags": []
},
"execution_count": 290
}
]
},
{
"metadata": {
"id": "TItXXK4GQw61",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 102
},
"outputId": "bbed17b3-2468-4e2f-9ad6-68c7a966a84b"
},
"cell_type": "code",
"source": [
"# Массив предсказаний\n",
"pred_test[:5]"
],
"execution_count": 291,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([[8396.412],\n",
" [8191.629],\n",
" [9359.061],\n",
" [8225.718],\n",
" [9147.268]], dtype=float32)"
]
},
"metadata": {
"tags": []
},
"execution_count": 291
}
]
},
{
"metadata": {
"id": "GCATQr3eRbT4",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 119
},
"outputId": "7b69cf8e-43a8-4b00-c5e6-b8b5c067dde6"
},
"cell_type": "code",
"source": [
"# Массив реальных курсов\n",
"real_test = np.exp(y_test)-1\n",
"real_test[:5]"
],
"execution_count": 292,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0 8188.959961\n",
"1 8300.820313\n",
"2 8876.080078\n",
"3 8934.719727\n",
"4 8822.360352\n",
"Name: y_test, dtype: float64"
]
},
"metadata": {
"tags": []
},
"execution_count": 292
}
]
},
{
"metadata": {
"id": "Jx5KjS2adInT",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 322
},
"outputId": "9b6683b5-2ebe-44ce-e626-a838037c7431"
},
"cell_type": "code",
"source": [
"# График предсказаний и реальных курсов\n",
"plt.figure(figsize=(12,5))\n",
"plt.plot(pred_test) # синий график\n",
"plt.plot(real_test) # зелёный график\n",
"plt.show()"
],
"execution_count": 293,
"outputs": [
{
"output_type": "display_data",
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAs8AAAExCAYAAACDAk9DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7fc1441b9d30>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"metadata": {
"id": "owZxZCKi4dfJ",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"Удивительно, что огибающая прогноза повторяет реальное состояние, возможно здесь есть какой то подвох."
]
},
{
"metadata": {
"id": "yMtsc-NlWs3N",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "b7c3432d-162a-4c9d-a35d-7d848326bf93"
},
"cell_type": "code",
"source": [
"lr = 5e-4\n",
"lr"
],
"execution_count": 294,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0.0005"
]
},
"metadata": {
"tags": []
},
"execution_count": 294
}
]
},
{
"metadata": {
"id": "5Gq3Mm-BPoZW",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 187
},
"outputId": "8a61c3e3-cb97-4dc4-e4d7-6ffbaaffc8db"
},
"cell_type": "code",
"source": [
"# Проведём ещё одно обучение\n",
"m.fit(lr, 3, cycle_len=1, cycle_mult=2, metrics=[metrics])"
],
"execution_count": 295,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "455b5d5231e74825a4c0746c13482f81",
"version_minor": 0,
"version_major": 2
},
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=7), HTML(value='')))"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
""
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
"epoch trn_loss val_loss metrics \n",
" 0 0.089974 0.031286 0.937871 \n",
" 25%|██▍ | 563/2265 [00:10<00:32, 51.60it/s, loss=0.105]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 1 0.095459 0.023066 0.896204 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0903]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 2 0.0629 0.018561 0.947114 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0579]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 3 0.075646 0.027587 0.930926 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0727]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 4 0.105853 0.025546 0.876446 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0968]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 5 0.086283 0.01173 0.94768 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0782]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 6 0.037318 0.010994 0.964631 \n"
],
"name": "stdout"
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[array([0.01099]), 0.9646313619824637]"
]
},
"metadata": {
"tags": []
},
"execution_count": 295
}
]
},
{
"metadata": {
"id": "TgnFJLWTtrAl",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "79fdb9e1-a354-4c03-ecf4-6fec3b466d51"
},
"cell_type": "code",
"source": [
"# Получим предсказания тестового набора\n",
"pred_test=m.predict(True)\n",
"pred_test = np.exp(pred_test)-1\n",
"'Ok'"
],
"execution_count": 298,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Ok'"
]
},
"metadata": {
"tags": []
},
"execution_count": 298
}
]
},
{
"metadata": {
"id": "5wN5cHA2ByJi",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 322
},
"outputId": "08056698-1787-4ffe-f157-406e0c8f7722"
},
"cell_type": "code",
"source": [
"# График предсказаний и реальных курсов\n",
"plt.figure(figsize=(12,5))\n",
"plt.plot(pred_test) # синий график\n",
"plt.plot(real_test) # зелёный график\n",
"plt.show()"
],
"execution_count": 297,
"outputs": [
{
"output_type": "display_data",
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAs8AAAExCAYAAACDAk9DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7fc1441c85f8>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"metadata": {
"id": "12WLy61qvoXU",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 187
},
"outputId": "f0de94ab-8bcf-49f4-e0b7-dcebaa10b291"
},
"cell_type": "code",
"source": [
"# Последний раз обучим\n",
"m.fit(lr, 3, cycle_len=1, cycle_mult=2, metrics=[metrics])"
],
"execution_count": 299,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "13fad2ce0e2d4e3d841d65b9c2b8491c",
"version_minor": 0,
"version_major": 2
},
"text/plain": [
"HBox(children=(IntProgress(value=0, description='Epoch', max=7), HTML(value='')))"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
""
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
"epoch trn_loss val_loss metrics \n",
" 0 0.038255 0.016298 0.950396 \n",
" 25%|██▍ | 560/2265 [00:10<00:32, 52.92it/s, loss=0.0852]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 1 0.039803 0.013391 0.931531 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0381]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 2 0.056948 0.012298 0.963247 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0522]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 3 0.055142 0.017796 0.939528 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0521]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 4 0.047892 0.014037 0.960652 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0443]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 5 0.026014 0.013736 0.941851 \n",
" 0%| | 0/2265 [00:00<?, ?it/s, loss=0.0247]"
],
"name": "stdout"
},
{
"output_type": "stream",
"text": [
" 6 0.042907 0.009572 0.962044 \n"
],
"name": "stdout"
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"[array([0.00957]), 0.9620442877828548]"
]
},
"metadata": {
"tags": []
},
"execution_count": 299
}
]
},
{
"metadata": {
"id": "sRaRV33Mw9xn",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "fcc4f7f0-e790-4ca3-b34c-fbf2e38e48e4"
},
"cell_type": "code",
"source": [
"# Получим предсказания тестового набора\n",
"pred_test=m.predict(True)\n",
"pred_test = np.exp(pred_test)-1\n",
"'Ok'"
],
"execution_count": 300,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Ok'"
]
},
"metadata": {
"tags": []
},
"execution_count": 300
}
]
},
{
"metadata": {
"id": "3OEjGo2dxBhp",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 320
},
"outputId": "9ecfdee0-f42a-420e-d347-746bb7eddbfa"
},
"cell_type": "code",
"source": [
"# График предсказаний и реальных курсов\n",
"plt.figure(figsize=(12,5))\n",
"plt.plot(pred_test) # синий график\n",
"plt.plot(real_test) # зелёный график\n",
"plt.show()"
],
"execution_count": 301,
"outputs": [
{
"output_type": "display_data",
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAs8AAAEvCAYAAAC63iyoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7fc144697e80>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"metadata": {
"id": "c3PBM7gI4FTn",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"Наглядно видно, чем больший срок прогноза, тем большее расхождение с реальным состоянием."
]
},
{
"metadata": {
"id": "iSuJg7JRZQLy",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "d0e7f412-d615-4a15-b33d-0c25badb1c57"
},
"cell_type": "code",
"source": [
"# Сохранить обученную(?) модель\n",
"m.save('probe2')\n",
"!ls '{PATH}models'"
],
"execution_count": 306,
"outputs": [
{
"output_type": "stream",
"text": [
"probe2.h5 tmp.h5\r\n"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "gHKiWd1aZ5Mq",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Скачать модель\n",
"files.download(f'{PATH}models/probe2.h5')"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "WYvJHxDkDQdU",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"### Использование модели"
]
},
{
"metadata": {
"id": "GETNFMMh0yS1",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"\n",
"\n",
"---\n",
"\n",
"Результаты способа, приведённого ниже некорректны, так как функция **proc_df()** (внутри пользовательской функции **data_probe2**) подготовки данных для модели уже по иному выдаёт значения, нежели при обучении...\n",
"В связи с этим, пока что только можно скачать и подготовить новый датасет до сегодняшнего дня (либо несколько последних дней не включать в набор), и заного обучать модель, что бы она прогнозировала завтрашний курс закрытия. "
]
},
{
"metadata": {
"id": "OobsSTsyau42",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "618cf358-cb93-449b-e7db-c55572f50db3"
},
"cell_type": "code",
"source": [
"# Проверить наличие модели (probe2.h5)\n",
"!ls '{PATH}models/'"
],
"execution_count": 308,
"outputs": [
{
"output_type": "stream",
"text": [
"probe2.h5 tmp.h5\r\n"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "3ZX2ospqa-nV",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Закачать модель (probe2.h5)\n",
"uploaded = files.upload()\n",
"!ls"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "Zwo0Gk3ebq_I",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Скопировать модель в рабочий каталог\n",
"!mv probe2.h5 '{PATH}models/'\n",
"!ls '{PATH}models/'"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "y7E_Rxh4fgmP",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Как подготовить модель для загрузки из файла пока вопрос"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "mlote082iUli",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Загрузить модель из файла\n",
"m.load('probe2')"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "J6b0gY7MFfqL",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 510
},
"outputId": "45f6188e-758b-44c4-ddc8-a0de62576546"
},
"cell_type": "code",
"source": [
"# Структура модели\n",
"m.summary"
],
"execution_count": 309,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"<bound method StructuredLearner.summary of MixedInputModel(\n",
" (embs): ModuleList(\n",
" (0): Embedding(10, 5)\n",
" (1): Embedding(13, 7)\n",
" (2): Embedding(54, 27)\n",
" (3): Embedding(8, 4)\n",
" (4): Embedding(3, 2)\n",
" (5): Embedding(3, 2)\n",
" (6): Embedding(3, 2)\n",
" (7): Embedding(3, 2)\n",
" (8): Embedding(3, 2)\n",
" (9): Embedding(3, 2)\n",
" )\n",
" (lins): ModuleList(\n",
" (0): Linear(in_features=61, out_features=1000, bias=True)\n",
" (1): Linear(in_features=1000, out_features=500, bias=True)\n",
" )\n",
" (bns): ModuleList(\n",
" (0): BatchNorm1d(1000, eps=1e-05, momentum=0.1, affine=True)\n",
" (1): BatchNorm1d(500, eps=1e-05, momentum=0.1, affine=True)\n",
" )\n",
" (outp): Linear(in_features=500, out_features=1, bias=True)\n",
" (emb_drop): Dropout(p=0.04)\n",
" (drops): ModuleList(\n",
" (0): Dropout(p=0.001)\n",
" (1): Dropout(p=0.01)\n",
" )\n",
" (bn): BatchNorm1d(6, eps=1e-05, momentum=0.1, affine=True)\n",
")>"
]
},
"metadata": {
"tags": []
},
"execution_count": 309
}
]
},
{
"metadata": {
"id": "ukJD8Qj3gVX_",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"**Прогнозы на тестовых данных**"
]
},
{
"metadata": {
"id": "YnQ_Ml7SwZ-l",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Функция предсказания модели структурных данных\n",
"def predict_struct(pand, cat, cont, m):\n",
" ''' Predicting a structural data model\n",
" \n",
" Args:\n",
" pand : pandas DataFrame\n",
" cat : numpy names of categorical data\n",
" cont : numpy names of continuous data\n",
" m : fastai model\n",
" '''\n",
" cat_values = pand[cat]\n",
" cont_values = pand[cont]\n",
" m.model.eval()\n",
" cat_values = to_gpu(V(T(np.array(cat_values))))\n",
" cont_values = to_gpu(V(T(np.array(cont_values))))\n",
" return to_np(m.model(cat_values, cont_values))"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "aZCZWhhI6x3F",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "e5e7d0f4-7245-4ec5-f696-7bea9d5222b2"
},
"cell_type": "code",
"source": [
"# Прогноз последнего дня тестового набора\n",
"result = predict_struct(df_test[-1:], cat_bin, contin_vars, m)\n",
"np.exp(result)-1"
],
"execution_count": 311,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([[13035.467]], dtype=float32)"
]
},
"metadata": {
"tags": []
},
"execution_count": 311
}
]
},
{
"metadata": {
"id": "3PdUZkf-9ni9",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 105
},
"outputId": "3e09bd95-a62d-487b-f1b1-f1692d705dad"
},
"cell_type": "code",
"source": [
"df_test[-1:]"
],
"execution_count": 312,
"outputs": [
{
"output_type": "execute_result",
"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>Year</th>\n",
" <th>Month</th>\n",
" <th>Week</th>\n",
" <th>Dayofweek</th>\n",
" <th>Is_month_end</th>\n",
" <th>Is_month_start</th>\n",
" <th>Is_quarter_end</th>\n",
" <th>Is_quarter_start</th>\n",
" <th>Is_year_end</th>\n",
" <th>Is_year_start</th>\n",
" <th>Open</th>\n",
" <th>High</th>\n",
" <th>Low</th>\n",
" <th>Close</th>\n",
" <th>Volume</th>\n",
" <th>Elapsed</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>98</th>\n",
" <td>8</td>\n",
" <td>7</td>\n",
" <td>30</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2.512239</td>\n",
" <td>2.662915</td>\n",
" <td>2.679441</td>\n",
" <td>2.764681</td>\n",
" <td>2.516219</td>\n",
" <td>1.852008</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Year Month Week Dayofweek Is_month_end Is_month_start \\\n",
"98 8 7 30 0 0 0 \n",
"\n",
" Is_quarter_end Is_quarter_start Is_year_end Is_year_start Open \\\n",
"98 0 0 0 0 2.512239 \n",
"\n",
" High Low Close Volume Elapsed \n",
"98 2.662915 2.679441 2.764681 2.516219 1.852008 "
]
},
"metadata": {
"tags": []
},
"execution_count": 312
}
]
},
{
"metadata": {
"id": "wG3Etln6glhd",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"\n",
"\n",
"---\n",
"\n",
"\n",
"**Прогноз на новых данных**"
]
},
{
"metadata": {
"id": "pmNQHn4dCr6h",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"Скачаем данные за 3 летних месяца:\n",
"\n",
"https://finance.yahoo.com/quote/BTC-USD/history?period1=1527800400&period2=1533070800&interval=1d&filter=history&frequency=1d"
]
},
{
"metadata": {
"id": "vYWRdrlMCrIT",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Загрузка исторических данных за 3 летних месяца (файл: BTC-USD_3m.csv)\n",
"# Скачать здесь: https://drive.google.com/file/d/17uIojBleJHAQAoxIrXYsDrKShyZtPz3X/view?usp=sharing\n",
"uploaded = files.upload()"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "qV7eZLBVDt8n",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "732232ea-4fa9-4165-8c80-73716c52eb79"
},
"cell_type": "code",
"source": [
"# Проверим\n",
"!ls"
],
"execution_count": 313,
"outputs": [
{
"output_type": "stream",
"text": [
"btc BTC-USD_3m (1).csv BTC-USD_3m.csv datalab lr_plot.png\r\n"
],
"name": "stdout"
}
]
},
{
"metadata": {
"id": "w6_vC2pQDtEX",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "2b330248-ac33-411a-a2c5-fbd2d0371fd8"
},
"cell_type": "code",
"source": [
"# Чтение данных\n",
"data3 = pd.read_csv('BTC-USD_3m.csv', error_bad_lines=False)[::1]\n",
"data3.shape"
],
"execution_count": 323,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(62, 7)"
]
},
"metadata": {
"tags": []
},
"execution_count": 323
}
]
},
{
"metadata": {
"id": "5lR3FH_5Edg7",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 204
},
"outputId": "8c01db5c-e242-4d9c-cf86-a50dc299050f"
},
"cell_type": "code",
"source": [
"data3.head()"
],
"execution_count": 324,
"outputs": [
{
"output_type": "execute_result",
"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>Date</th>\n",
" <th>Open</th>\n",
" <th>High</th>\n",
" <th>Low</th>\n",
" <th>Close</th>\n",
" <th>Adj Close</th>\n",
" <th>Volume</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>2018-05-30</td>\n",
" <td>7393.020020</td>\n",
" <td>7608.870117</td>\n",
" <td>7349.520020</td>\n",
" <td>7502.149902</td>\n",
" <td>7502.149902</td>\n",
" <td>458223966</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>2018-05-31</td>\n",
" <td>7501.740234</td>\n",
" <td>7614.660156</td>\n",
" <td>7370.270020</td>\n",
" <td>7530.549805</td>\n",
" <td>7530.549805</td>\n",
" <td>458687659</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2018-06-01</td>\n",
" <td>7530.549805</td>\n",
" <td>7697.339844</td>\n",
" <td>7467.790039</td>\n",
" <td>7643.259766</td>\n",
" <td>7643.259766</td>\n",
" <td>362414878</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>2018-06-02</td>\n",
" <td>7643.259766</td>\n",
" <td>7774.959961</td>\n",
" <td>7606.759766</td>\n",
" <td>7719.750000</td>\n",
" <td>7719.750000</td>\n",
" <td>332313005</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>2018-06-03</td>\n",
" <td>7719.729980</td>\n",
" <td>7760.729980</td>\n",
" <td>7469.209961</td>\n",
" <td>7503.200195</td>\n",
" <td>7503.200195</td>\n",
" <td>427448262</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Date Open High Low Close \\\n",
"0 2018-05-30 7393.020020 7608.870117 7349.520020 7502.149902 \n",
"1 2018-05-31 7501.740234 7614.660156 7370.270020 7530.549805 \n",
"2 2018-06-01 7530.549805 7697.339844 7467.790039 7643.259766 \n",
"3 2018-06-02 7643.259766 7774.959961 7606.759766 7719.750000 \n",
"4 2018-06-03 7719.729980 7760.729980 7469.209961 7503.200195 \n",
"\n",
" Adj Close Volume \n",
"0 7502.149902 458223966 \n",
"1 7530.549805 458687659 \n",
"2 7643.259766 362414878 \n",
"3 7719.750000 332313005 \n",
"4 7503.200195 427448262 "
]
},
"metadata": {
"tags": []
},
"execution_count": 324
}
]
},
{
"metadata": {
"id": "61EKGDE9EvmC",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
"# Функция подготовки данных для модели probe2\n",
"def data_probe2(arg):\n",
" '''Data preparation function for model probe2\n",
" Args:\n",
" pandas DataFrame taken from a csv file downloaded from the site https://finance.yahoo.com/quote/BTC-USD/history\n",
" Returns:\n",
" data : prepared for the model DataFrame\n",
" categ : list of categorical data names\n",
" contin : list of continuous data names\n",
" '''\n",
" # copy DataFrame\n",
" data = arg.copy()\n",
" # data supplement\n",
" add_datepart(data, \"Date\", drop=False)\n",
" # removal of excess\n",
" data.drop('Date', axis=1, inplace=True)\n",
" data.drop('Day', axis=1, inplace=True)\n",
" data.drop('Dayofyear', axis=1, inplace=True)\n",
" #data.drop('Adj Close', axis=1, inplace=True)\n",
" # bringing categorical data\n",
" data['Year'] -= np.min(data['Year'])\n",
" data.Is_month_end = data.Is_month_end.astype('int')\n",
" data.Is_month_start = data.Is_month_start.astype('int')\n",
" data.Is_quarter_end = data.Is_quarter_end.astype('int')\n",
" data.Is_quarter_start = data.Is_quarter_start.astype('int')\n",
" data.Is_year_end = data.Is_year_end.astype('int')\n",
" data.Is_year_start = data.Is_year_start.astype('int')\n",
" # list of categorical and continuous data\n",
" categ = ['Year', 'Month', 'Week', 'Dayofweek', 'Is_month_end', 'Is_month_start', 'Is_quarter_end', 'Is_quarter_start', 'Is_year_end', 'Is_year_start']\n",
" contin = ['Open', 'High', 'Low', 'Close', 'Volume', 'Elapsed']\n",
" # preparing data for the model\n",
" df3, _, _, _ = proc_df(data, 'Adj Close', ignore_flds=cat_bin, do_scale=True,)\n",
" # data prepared for the forecast close price for the day ahead\n",
" return df3, categ, contin"
],
"execution_count": 0,
"outputs": []
},
{
"metadata": {
"id": "87Jp4JPRGRtf",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 204
},
"outputId": "7dbc8aa7-9622-46ad-b240-8784b913eca6"
},
"cell_type": "code",
"source": [
"# Подготовка данных\n",
"df3, categ, contin = data_probe2(data3)\n",
"categ, contin"
],
"execution_count": 325,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(['Year',\n",
" 'Month',\n",
" 'Week',\n",
" 'Dayofweek',\n",
" 'Is_month_end',\n",
" 'Is_month_start',\n",
" 'Is_quarter_end',\n",
" 'Is_quarter_start',\n",
" 'Is_year_end',\n",
" 'Is_year_start'],\n",
" ['Open', 'High', 'Low', 'Close', 'Volume', 'Elapsed'])"
]
},
"metadata": {
"tags": []
},
"execution_count": 325
}
]
},
{
"metadata": {
"id": "oMmIWYzVaJay",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "0d3ae3e4-3dbe-46ea-ffa8-303a9f90f443"
},
"cell_type": "code",
"source": [
"# Предсказание по полученному набору данных\n",
"result = predict_struct(df3, cat_bin, contin, m)\n",
"result = np.exp(result)-1\n",
"len(result)"
],
"execution_count": 328,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"62"
]
},
"metadata": {
"tags": []
},
"execution_count": 328
}
]
},
{
"metadata": {
"id": "ZOsWUHBcfdbJ",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 714
},
"outputId": "fef1a578-968e-4aa4-acc2-367c1c8d8aff"
},
"cell_type": "code",
"source": [
"# Показать прогнозы\n",
"result"
],
"execution_count": 329,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([[ 0.93905],\n",
" [ 1.18122],\n",
" [ 3.11828],\n",
" [ 4.2434 ],\n",
" [ 3.65577],\n",
" [ 4.23007],\n",
" [ 5.8391 ],\n",
" [ 7.21232],\n",
" [ 7.5835 ],\n",
" [ 8.00627],\n",
" [ 1.99735],\n",
" [ 0.93621],\n",
" [ 0.64297],\n",
" [ 0.34054],\n",
" [ 0.37219],\n",
" [ 0.5413 ],\n",
" [ 0.44963],\n",
" [ 0.64018],\n",
" [ 0.70153],\n",
" [ 2.09366],\n",
" ...,\n",
" [ 5.36852],\n",
" [ 5.2411 ],\n",
" [ 6.53773],\n",
" [ 7.60264],\n",
" [ 8.57682],\n",
" [ 14.48204],\n",
" [ 51.8365 ],\n",
" [ 79.13386],\n",
" [ 83.88987],\n",
" [ 93.14674],\n",
" [122.9342 ],\n",
" [139.07036],\n",
" [272.11548],\n",
" [448.46072],\n",
" [444.41086],\n",
" [457.94705],\n",
" [852.7607 ],\n",
" [941.20844],\n",
" [606.4219 ],\n",
" [481.21613]], dtype=float32)"
]
},
"metadata": {
"tags": []
},
"execution_count": 329
}
]
},
{
"metadata": {
"id": "BfRE7xzufslK",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 68
},
"outputId": "110d2b62-a7f8-4c31-e23c-d9d9d03ba563"
},
"cell_type": "code",
"source": [
"# Показать реальные курсы\n",
"data3['Close'].values[-20:]"
],
"execution_count": 320,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([6253.6001 , 6229.83008, 6268.75 , 6364.25977, 6740.5498 , 7326.7002 , 7383.39014, 7477.5 ,\n",
" 7333.93018, 7405.3999 , 7398.64014, 7718. , 8395.82031, 8170.22998, 7937.25 , 8182.89014,\n",
" 8230.87012, 8216.78027, 8176.06006, 7735.2998 ])"
]
},
"metadata": {
"tags": []
},
"execution_count": 320
}
]
},
{
"metadata": {
"id": "pysLfIhdkGWr",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 85
},
"outputId": "3ff3d94d-2c48-481f-fefa-1ac3aa03a43a"
},
"cell_type": "code",
"source": [
"real_test[-30:].values"
],
"execution_count": 321,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([6087.39014, 6140.56982, 5870.27978, 6202.7998 , 6384.37988, 6338.04004, 6614.66016, 6508.58008,\n",
" 6589.06006, 6533.81006, 6601.02002, 6757.08008, 6706.37988, 6667.83984, 6305.8501 , 6393.35986,\n",
" 6252.6001 , 6228.83008, 6267.75 , 6363.25977, 6739.5498 , 7325.70019, 7382.39014, 7476.5 ,\n",
" 7332.93018, 7404.3999 , 7397.64014, 7717. , 8394.82031, 8169.22998])"
]
},
"metadata": {
"tags": []
},
"execution_count": 321
}
]
},
{
"metadata": {
"id": "-dh6SGijcIWa",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 320
},
"outputId": "c5027b0c-3a26-4545-a850-57b322abb7a2"
},
"cell_type": "code",
"source": [
"# График предсказаний и реальных курсов\n",
"plt.figure(figsize=(12,5))\n",
"plt.plot(result[:-7]) # синий график прогнозы 3 месяца\n",
"plt.plot((data3['Close'].values)[:-6]) # зелёный график 3 месяца\n",
"plt.plot((pred_test[-57:])) # красный график прогнозы теста\n",
"plt.plot((real_test[-57:]).values) # синий график реальный\n",
"plt.show()"
],
"execution_count": 330,
"outputs": [
{
"output_type": "display_data",
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAs8AAAEvCAYAAAC63iyoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7fc14468b240>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"metadata": {
"id": "P7ufRElJrCsv",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 229
},
"outputId": "29a9c97c-d1b2-44db-dd5b-9f0e2d1bf083"
},
"cell_type": "code",
"source": [
"df3[:-7].head()"
],
"execution_count": 336,
"outputs": [
{
"output_type": "execute_result",
"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>Year</th>\n",
" <th>Month</th>\n",
" <th>Week</th>\n",
" <th>Dayofweek</th>\n",
" <th>Is_month_end</th>\n",
" <th>Is_month_start</th>\n",
" <th>Is_quarter_end</th>\n",
" <th>Is_quarter_start</th>\n",
" <th>Is_year_end</th>\n",
" <th>Is_year_start</th>\n",
" <th>Open</th>\n",
" <th>High</th>\n",
" <th>Low</th>\n",
" <th>Close</th>\n",
" <th>Volume</th>\n",
" <th>Elapsed</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>5</td>\n",
" <td>22</td>\n",
" <td>2</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.639755</td>\n",
" <td>0.725700</td>\n",
" <td>0.821500</td>\n",
" <td>0.786839</td>\n",
" <td>-0.026645</td>\n",
" <td>-1.704336</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>0</td>\n",
" <td>5</td>\n",
" <td>22</td>\n",
" <td>3</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.800030</td>\n",
" <td>0.734174</td>\n",
" <td>0.851942</td>\n",
" <td>0.828400</td>\n",
" <td>-0.024295</td>\n",
" <td>-1.648456</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>0</td>\n",
" <td>6</td>\n",
" <td>22</td>\n",
" <td>4</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0.842501</td>\n",
" <td>0.855180</td>\n",
" <td>0.995011</td>\n",
" <td>0.993343</td>\n",
" <td>-0.512345</td>\n",
" <td>-1.592576</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>0</td>\n",
" <td>6</td>\n",
" <td>22</td>\n",
" <td>5</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1.008658</td>\n",
" <td>0.968780</td>\n",
" <td>1.198889</td>\n",
" <td>1.105280</td>\n",
" <td>-0.664944</td>\n",
" <td>-1.536697</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>0</td>\n",
" <td>6</td>\n",
" <td>22</td>\n",
" <td>6</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1.121390</td>\n",
" <td>0.947954</td>\n",
" <td>0.997094</td>\n",
" <td>0.788376</td>\n",
" <td>-0.182661</td>\n",
" <td>-1.480817</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Year Month Week Dayofweek Is_month_end Is_month_start Is_quarter_end \\\n",
"0 0 5 22 2 0 0 0 \n",
"1 0 5 22 3 1 0 0 \n",
"2 0 6 22 4 0 1 0 \n",
"3 0 6 22 5 0 0 0 \n",
"4 0 6 22 6 0 0 0 \n",
"\n",
" Is_quarter_start Is_year_end Is_year_start Open High Low \\\n",
"0 0 0 0 0.639755 0.725700 0.821500 \n",
"1 0 0 0 0.800030 0.734174 0.851942 \n",
"2 0 0 0 0.842501 0.855180 0.995011 \n",
"3 0 0 0 1.008658 0.968780 1.198889 \n",
"4 0 0 0 1.121390 0.947954 0.997094 \n",
"\n",
" Close Volume Elapsed \n",
"0 0.786839 -0.026645 -1.704336 \n",
"1 0.828400 -0.024295 -1.648456 \n",
"2 0.993343 -0.512345 -1.592576 \n",
"3 1.105280 -0.664944 -1.536697 \n",
"4 0.788376 -0.182661 -1.480817 "
]
},
"metadata": {
"tags": []
},
"execution_count": 336
}
]
},
{
"metadata": {
"id": "ASEngZ8szrCf",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 229
},
"outputId": "00894617-bbe0-4709-bd41-2473738a4c55"
},
"cell_type": "code",
"source": [
"df_test[-55:].head()"
],
"execution_count": 340,
"outputs": [
{
"output_type": "execute_result",
"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>Year</th>\n",
" <th>Month</th>\n",
" <th>Week</th>\n",
" <th>Dayofweek</th>\n",
" <th>Is_month_end</th>\n",
" <th>Is_month_start</th>\n",
" <th>Is_quarter_end</th>\n",
" <th>Is_quarter_start</th>\n",
" <th>Is_year_end</th>\n",
" <th>Is_year_start</th>\n",
" <th>Open</th>\n",
" <th>High</th>\n",
" <th>Low</th>\n",
" <th>Close</th>\n",
" <th>Volume</th>\n",
" <th>Elapsed</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>44</th>\n",
" <td>8</td>\n",
" <td>5</td>\n",
" <td>22</td>\n",
" <td>2</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2.389089</td>\n",
" <td>2.345354</td>\n",
" <td>2.541481</td>\n",
" <td>2.426435</td>\n",
" <td>0.807293</td>\n",
" <td>1.785933</td>\n",
" </tr>\n",
" <tr>\n",
" <th>45</th>\n",
" <td>8</td>\n",
" <td>5</td>\n",
" <td>22</td>\n",
" <td>3</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2.430288</td>\n",
" <td>2.347451</td>\n",
" <td>2.549836</td>\n",
" <td>2.437184</td>\n",
" <td>0.808412</td>\n",
" <td>1.787157</td>\n",
" </tr>\n",
" <tr>\n",
" <th>46</th>\n",
" <td>8</td>\n",
" <td>6</td>\n",
" <td>22</td>\n",
" <td>4</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2.441205</td>\n",
" <td>2.377395</td>\n",
" <td>2.589101</td>\n",
" <td>2.479844</td>\n",
" <td>0.576145</td>\n",
" <td>1.788381</td>\n",
" </tr>\n",
" <tr>\n",
" <th>47</th>\n",
" <td>8</td>\n",
" <td>6</td>\n",
" <td>22</td>\n",
" <td>5</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2.483916</td>\n",
" <td>2.405506</td>\n",
" <td>2.645056</td>\n",
" <td>2.508795</td>\n",
" <td>0.503521</td>\n",
" <td>1.789604</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48</th>\n",
" <td>8</td>\n",
" <td>6</td>\n",
" <td>22</td>\n",
" <td>6</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2.512895</td>\n",
" <td>2.400352</td>\n",
" <td>2.589673</td>\n",
" <td>2.426832</td>\n",
" <td>0.733044</td>\n",
" <td>1.790828</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Year Month Week Dayofweek Is_month_end Is_month_start \\\n",
"44 8 5 22 2 0 0 \n",
"45 8 5 22 3 1 0 \n",
"46 8 6 22 4 0 1 \n",
"47 8 6 22 5 0 0 \n",
"48 8 6 22 6 0 0 \n",
"\n",
" Is_quarter_end Is_quarter_start Is_year_end Is_year_start Open \\\n",
"44 0 0 0 0 2.389089 \n",
"45 0 0 0 0 2.430288 \n",
"46 0 0 0 0 2.441205 \n",
"47 0 0 0 0 2.483916 \n",
"48 0 0 0 0 2.512895 \n",
"\n",
" High Low Close Volume Elapsed \n",
"44 2.345354 2.541481 2.426435 0.807293 1.785933 \n",
"45 2.347451 2.549836 2.437184 0.808412 1.787157 \n",
"46 2.377395 2.589101 2.479844 0.576145 1.788381 \n",
"47 2.405506 2.645056 2.508795 0.503521 1.789604 \n",
"48 2.400352 2.589673 2.426832 0.733044 1.790828 "
]
},
"metadata": {
"tags": []
},
"execution_count": 340
}
]
},
{
"metadata": {
"id": "x_gSI8C-g17W",
"colab_type": "text"
},
"cell_type": "markdown",
"source": [
"Теперь видно, что по разному подготовились данные для обучения и для прогнозирования."
]
},
{
"metadata": {
"id": "CLW17s6LhBnV",
"colab_type": "code",
"colab": {}
},
"cell_type": "code",
"source": [
""
],
"execution_count": 0,
"outputs": []
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment