logistic_regression_pytorch.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Логистическая регрессия с помощью PyTorch"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Содержание\n",
    "\n",
    "1. Бинарная классификация с двумя параметрами на iris.csv  \n",
    "2. Бинарная классификация с четырьмя параметрами на iris.csv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Бинарная классификация с двумя параметрами на iris.csv "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Предобработка данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt # для визуализации\n",
    "import os # для обработки пути к данным как на Unix-системах, так и на Windows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(42) # фиксирование псевдорандомного генератора"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATA_DIR = \"data\" # путь к данным"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sepal_length</th>\n",
       "      <th>sepal_width</th>\n",
       "      <th>petal_length</th>\n",
       "      <th>petal_width</th>\n",
       "      <th>species</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6.0</td>\n",
       "      <td>2.7</td>\n",
       "      <td>5.1</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.5</td>\n",
       "      <td>2.3</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.9</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.8</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.8</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.1</td>\n",
       "      <td>3.8</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sepal_length  sepal_width  petal_length  petal_width  species\n",
       "0           6.0          2.7           5.1          1.6        1\n",
       "1           5.5          2.3           4.0          1.3        1\n",
       "2           5.9          3.2           4.8          1.8        1\n",
       "3           4.8          3.0           1.4          0.3        0\n",
       "4           5.1          3.8           1.9          0.4        0"
      ]
     },
     "execution_count": 281,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_iris = pd.read_csv(os.path.join(DATA_DIR, 'iris.csv')) # прочитать данные из CSV\n",
    "df_iris = df_iris.iloc[:100] # взять только два вида ирисов для бинарной классификации\n",
    "df_iris = df_iris.sample(frac=1, random_state=42) # перемешать данные\n",
    "df_iris = df_iris.reset_index(drop=True) # сбросить индексы для красоты\n",
    "\n",
    "# можно было записать одной строкой\n",
    "# df_iris = pd.read_csv(os.path.join(DATA_DIR, 'iris.csv')).iloc[:100].sample(frac=1, random_state=42).reset_index(drop=True)\n",
    "\n",
    "df_iris['species'] = df_iris['species'].astype('category').cat.codes # заменить setosa и versicolor на 0 и 1\n",
    "df_iris.head() # вывести первые 5 рядов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train.shape: (67, 2)\n",
      "y_train.shape: (67,)\n",
      "X_test.shape: (33, 2)\n",
      "y_test.shape: (33,)\n"
     ]
    }
   ],
   "source": [
    "# тренировочный датасет\n",
    "X_train = df_iris[['sepal_length', 'sepal_width']][:67].values # два параметра - sepal_length и sepal_width\n",
    "y_train = df_iris['species'][:67].values\n",
    "\n",
    "# тестовый датасет\n",
    "X_test = df_iris[['sepal_length', 'sepal_width']][67:].values # два параметра - sepal_length и sepal_width\n",
    "y_test = df_iris['species'][67:].values\n",
    "\n",
    "print(\"X_train.shape: {}\".format(X_train.shape))\n",
    "print(\"y_train.shape: {}\".format(y_train.shape))\n",
    "print(\"X_test.shape: {}\".format(X_test.shape))\n",
    "print(\"y_test.shape: {}\".format(y_test.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Визуализация данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X_train[y_train == 0][:, 0], X_train[y_train == 0][:, 1], 'rx', label='Train - Setosa')\n",
    "\n",
    "plt.plot(X_train[y_train == 1][:, 0], X_train[y_train == 1][:, 1], 'bx', label='Train - Versicolor')\n",
    "\n",
    "plt.xlabel(\"Sepal length\") # надпись по оси X\n",
    "plt.ylabel(\"Sepal width\") # надпись по оси Y\n",
    "plt.legend() # отображение легенды\n",
    "plt.show() # отображение графика"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Создание и тренировка модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "########################################## TODO ########################################################"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Оценка модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "########################################## TODO ########################################################"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Визуализация работы модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 287,
   "metadata": {},
   "outputs": [],
   "source": [
    "h = 0.01\n",
    "x_min, x_max = X_train[:, 0].min()-1, X_train[:, 0].max()+1\n",
    "y_min, y_max = X_train[:, 1].min()-1, X_train[:, 1].max()+1\n",
    "xx, yy = np.meshgrid(np.arange(x_min, x_max, h),\n",
    "                     np.arange(y_min, y_max, h))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
    "Z = Z.reshape(xx.shape)\n",
    "plt.plot(X_train[y_train == 0][:, 0], X_train[y_train == 0][:, 1], 'rx', label='Train - Setosa')\n",
    "plt.plot(X_train[y_train == 1][:, 0], X_train[y_train == 1][:, 1], 'bx', label='Train - Versicolor')\n",
    "plt.plot(X_test[y_test == 0][:, 0], X_test[y_test == 0][:, 1], 'ro', label='Test - Setosa')\n",
    "plt.plot(X_test[y_test == 1][:, 0], X_test[y_test == 1][:, 1], 'bo', label='Test - Versicolor')\n",
    "plt.contourf(xx, yy, Z, cmap=plt.cm.RdBu, alpha=0.1) # цветовые области\n",
    "\n",
    "plt.xlabel(\"Sepal length\") # надпись по оси X\n",
    "plt.ylabel(\"Sepal width\") # надпись по оси Y\n",
    "plt.legend() # отображение легенды\n",
    "plt.show() # отображение графика"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Бинарная классификация с четырьмя параметрами на iris.csv "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Предобработка данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 289,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os # для обработки пути к данным как на Unix-системах, так и на Windows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(42) # фиксирование псевдорандомного генератора"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATA_DIR = \"data\" # путь к данным"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sepal_length</th>\n",
       "      <th>sepal_width</th>\n",
       "      <th>petal_length</th>\n",
       "      <th>petal_width</th>\n",
       "      <th>species</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6.0</td>\n",
       "      <td>2.7</td>\n",
       "      <td>5.1</td>\n",
       "      <td>1.6</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.5</td>\n",
       "      <td>2.3</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.9</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.8</td>\n",
       "      <td>1.8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.8</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.1</td>\n",
       "      <td>3.8</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sepal_length  sepal_width  petal_length  petal_width  species\n",
       "0           6.0          2.7           5.1          1.6        1\n",
       "1           5.5          2.3           4.0          1.3        1\n",
       "2           5.9          3.2           4.8          1.8        1\n",
       "3           4.8          3.0           1.4          0.3        0\n",
       "4           5.1          3.8           1.9          0.4        0"
      ]
     },
     "execution_count": 292,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_iris = pd.read_csv(os.path.join(DATA_DIR, 'iris.csv')) # прочитать данные из CSV\n",
    "df_iris = df_iris.iloc[:100] # взять только два вида ирисов для бинарной классификации\n",
    "df_iris = df_iris.sample(frac=1, random_state=42) # перемешать данные\n",
    "df_iris = df_iris.reset_index(drop=True) # сбросить индексы для красоты\n",
    "\n",
    "# можно было записать одной строкой\n",
    "# df_iris = pd.read_csv(os.path.join(DATA_DIR, 'iris.csv')).iloc[:100].sample(frac=1, random_state=42).reset_index(drop=True)\n",
    "\n",
    "df_iris['species'] = df_iris['species'].astype('category').cat.codes # заменить setosa и versicolor на 0 и 1\n",
    "df_iris.head() # вывести первые 5 рядов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train.shape: (67, 4)\n",
      "y_train.shape: (67,)\n",
      "X_test.shape: (33, 4)\n",
      "y_test.shape: (33,)\n"
     ]
    }
   ],
   "source": [
    "# тренировочный датасет\n",
    "X_train = df_iris.drop(['species'], axis=1).iloc[:67].values # четыре параметра\n",
    "y_train = df_iris['species'][:67].values\n",
    "\n",
    "# тестовый датасет\n",
    "X_test = df_iris.drop(['species'], axis=1).iloc[67:].values # четыре параметра\n",
    "y_test = df_iris['species'][67:].values\n",
    "\n",
    "print(\"X_train.shape: {}\".format(X_train.shape))\n",
    "print(\"y_train.shape: {}\".format(y_train.shape))\n",
    "print(\"X_test.shape: {}\".format(X_test.shape))\n",
    "print(\"y_test.shape: {}\".format(y_test.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Создание и тренировка модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "########################################## TODO ########################################################"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Оценка модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "########################################## TODO ########################################################"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}