{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "## **Лабораторная работа 4 (Diamonds Prices)**" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "import pandas as pd\n", "df = pd.read_csv(\".//static//csv//Diamonds Prices2022.csv\")\n", "print(df.info)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Задача регрессии**\n", "\n", "#### **Бизнес-цель для задачи регрессии**\n", "\n", "Предсказать цену бриллианта на основе его характеристик\n", "\n", "Зачем это нужно?\n", "\n", "Для покупателей: Упрощение выбора бриллианта, основываясь на соотношении цена-качество.\n", "\n", "Для продавцов: Оптимизация ценовой политики и привлечение клиентов за счет точного ценообразования.\n", "\n", "Для оценщиков: Ускорение процесса оценки бриллиантов и снижение субъективности" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### **Достижимый уровень качества модели для задачи регрессии:**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Целевой уровень:** MAE < 300, R² > 0.97. \n", "\n", "**Основание:** это позволяет достичь приемлемой точности для предсказания цен бриллиантов с учетом их характеристик.\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Линейная регрессия**" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Лучшие гиперпараметры: {'preprocessing': MaxAbsScaler()}\n", "Cредняя абсолютная ошибка (MAE) = 731.547938145698\n", "R^2 = 0.9221982562588074\n", "Смещение: -1138.4103165293116\n", "Дисперсия: 4.4030722917880405e-13\n" ] } ], "source": [ "import pandas as pd\n", "from sklearn.model_selection import train_test_split, GridSearchCV\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.linear_model import LinearRegression\n", "from sklearn.preprocessing import StandardScaler, MinMaxScaler, MaxAbsScaler\n", "from sklearn.metrics import mean_absolute_error, r2_score\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.impute import SimpleImputer\n", "import sklearn.preprocessing as preproc\n", "import numpy as np\n", "\n", "df = pd.read_csv(\".//static//csv//Diamonds Prices2022.csv\")\n", "data = df[['carat', 'cut', 'color', 'clarity', 'depth', 'table', 'x', 'y', 'z', 'price']]\n", "\n", "# Преобразуем категориальные признаки в числовые через one-hot encoding\n", "data = pd.get_dummies(data, columns=['cut', 'color', 'clarity'], drop_first=True)\n", "\n", "# Определение входных и целевых переменных\n", "X = data.drop('price', axis=1) # признаки\n", "y = data['price'] # целевая переменная (цена)\n", "\n", "# Разделение данных на обучающую и тестовую выборки\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n", "\n", "# Подготовка пайплайна для линейной регрессии\n", "pipeline_lin_reg = Pipeline([\n", " ('preprocessing', ColumnTransformer([\n", " ('num', SimpleImputer(strategy='median'), X.columns) \n", " ])),\n", " ('model', LinearRegression()) \n", "])\n", "\n", "# Определение сетки гиперпараметров для поиска лучших значений\n", "param_grid = {\n", " 'preprocessing': [StandardScaler(), MinMaxScaler(), MaxAbsScaler(), None] # разные методы масштабирования\n", "}\n", "\n", "# Создание объекта GridSearchCV для поиска лучших гиперпараметров\n", "grid_search = GridSearchCV(pipeline_lin_reg, param_grid, cv=5, scoring='neg_root_mean_squared_error', n_jobs=-1)\n", "\n", "# Обучение модели с перебором гиперпараметров\n", "grid_search.fit(X_train, y_train)\n", "\n", "# Вывод лучших гиперпараметров\n", "print(\"Лучшие гиперпараметры: \", grid_search.best_params_)\n", "\n", "# Лучшая модель линейной регрессии\n", "best_model = grid_search.best_estimator_\n", "\n", "# Прогнозирование на тестовых данных\n", "y_pred = best_model.predict(X_test)\n", "\n", "# Оценка модели: Средняя абсолютная ошибка (MAE)\n", "mae = mean_absolute_error(y_test, y_pred)\n", "print(f'Cредняя абсолютная ошибка (MAE) = {mae}')\n", "\n", "# Оценка качества модели: R^2\n", "r2 = r2_score(y_test, y_pred)\n", "print(f'R^2 = {r2}')\n", "\n", "# Оценка дисперсии и смещения\n", "cv_results = grid_search.cv_results_\n", "mean_test_score = np.mean(cv_results['mean_test_score'])\n", "std_test_score = np.std(cv_results['mean_test_score'])\n", "\n", "print(f\"Смещение: {mean_test_score}\")\n", "print(f\"Дисперсия: {std_test_score}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Градиентный бустинг**" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Лучшие гиперпараметры: {'model__learning_rate': 0.1, 'model__max_depth': 7, 'model__n_estimators': 300, 'preprocessing': StandardScaler()}\n", "Cредняя абсолютная ошибка (MAE) = 282.34135362359893\n", "Смещение: -696.9996831010233\n", "Дисперсия: 4.393470605801296\n", "R^2 = 0.9790739156988512\n" ] } ], "source": [ "from sklearn.ensemble import GradientBoostingRegressor\n", "from sklearn.model_selection import cross_val_predict\n", "\n", "# Преобразование числовых данных (заполнение пустых значений медианой)\n", "num_imputer = SimpleImputer(strategy=\"median\")\n", "preprocessing_num = Pipeline([(\"imputer\", num_imputer)])\n", "\n", "# Общая предобработка (только числовые данные)\n", "preprocessing = ColumnTransformer(\n", " [(\"nums\", preprocessing_num, X_train.columns)]\n", ")\n", "\n", "# Конвейер\n", "pipeline_grad = Pipeline([\n", " ('preprocessing', preprocessing),\n", " ('model', GradientBoostingRegressor(random_state=42))\n", "])\n", "\n", "# Определение сетки гиперпараметров\n", "param_grid = {\n", " 'preprocessing': [StandardScaler(), preproc.MinMaxScaler(), preproc.MaxAbsScaler(), None],\n", " 'model__n_estimators': [100, 200, 300],\n", " 'model__learning_rate': [0.1, 0.2],\n", " 'model__max_depth': [3, 5, 7]\n", "}\n", "\n", "# Создание объекта GridSearchCV\n", "grid_search = GridSearchCV(pipeline_grad, param_grid, cv=2, scoring='neg_root_mean_squared_error', n_jobs=-1)\n", "\n", "# Обучение модели с перебором гиперпараметров\n", "grid_search.fit(X_train, y_train)\n", "\n", "# Вывод лучших гиперпараметров\n", "print(\"Лучшие гиперпараметры: \", grid_search.best_params_)\n", "\n", "# Лучшая модель градиентного бустинга\n", "best_model = grid_search.best_estimator_\n", "\n", "# Предсказания на тестовой выборке\n", "y_pred = best_model.predict(X_test)\n", "\n", "# Оценка качества\n", "print(f'Cредняя абсолютная ошибка (MAE) = {mean_absolute_error(y_test, y_pred)}')\n", "\n", "# Предсказания на кросс-валидации\n", "y_cv_pred = cross_val_predict(best_model, X_train, y_train, cv=3)\n", "\n", "# Оценка смещения и дисперсии\n", "cv_results = grid_search.cv_results_\n", "mean_test_score = cv_results['mean_test_score']\n", "std_test_score = cv_results['std_test_score']\n", "\n", "print(f\"Смещение: {mean_test_score.mean()}\")\n", "print(f\"Дисперсия: {std_test_score.mean()}\")\n", "\n", "# Коэффициент детерминации R^2\n", "print(f'R^2 = {r2_score(y_test, y_pred)}')\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Случаные леса**" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Лучшие гиперпараметры: {'model__max_depth': None, 'model__min_samples_split': 5, 'model__n_estimators': 300, 'preprocessing': None}\n", "Средняя абсолютная ошибка (MAE) = 292.2620729910037\n", "Смещение: -882.5346885343729\n", "Дисперсия: 11.135626719319792\n", "R^2 = 0.9743449800931261\n" ] } ], "source": [ "from sklearn.ensemble import RandomForestRegressor\n", "\n", "# Конвейер\n", "pipeline_forest = Pipeline([\n", " ('preprocessing', preprocessing),\n", " ('model', RandomForestRegressor(random_state=42))\n", "])\n", "\n", "# Определение сетки гиперпараметров\n", "param_grid = {\n", " 'preprocessing': [StandardScaler(), preproc.MinMaxScaler(), preproc.MaxAbsScaler(), None],\n", " 'model__n_estimators': [100, 200, 300], # Количество деревьев\n", " 'model__max_depth': [None, 5, 10], # Максимальная глубина деревьев\n", " 'model__min_samples_split': [2, 5], # Минимальное число образцов для разделения\n", "}\n", "\n", "# Создание объекта GridSearchCV\n", "grid_search = GridSearchCV(pipeline_forest, param_grid, cv=3, scoring='neg_root_mean_squared_error', n_jobs=-1)\n", "\n", "# Обучение модели с перебором гиперпараметров\n", "grid_search.fit(X_train, y_train)\n", "\n", "print(\"Лучшие гиперпараметры: \", grid_search.best_params_)\n", "\n", "# Лучшая модель случайного леса\n", "best_model = grid_search.best_estimator_\n", "\n", "y_pred = best_model.predict(X_test)\n", "\n", "print(f'Средняя абсолютная ошибка (MAE) = {mean_absolute_error(y_test, y_pred)}')\n", "\n", "# Получение предсказаний на кросс-валидации\n", "y_cv_pred = cross_val_predict(best_model, X_train, y_train, cv=3)\n", "\n", "cv_results = grid_search.cv_results_\n", "mean_test_score = cv_results['mean_test_score']\n", "std_test_score = cv_results['std_test_score']\n", "\n", "print(f\"Смещение: {mean_test_score.mean()}\")\n", "print(f\"Дисперсия: {std_test_score.mean()}\")\n", "\n", "print(f'R^2 = {r2_score(y_test, y_pred)}')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Вывод:**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Линейная регрессия:**\n", "\n", "MAE: 731.55 — наибольшая ошибка среди моделей.\n", "\n", "R²: 0.9222 — показывает, что модель объясняет 92.2% дисперсии, но хуже справляется с нелинейными зависимостями.\n", "\n", "Смещение: -1138.41 — модель недооценивает значения (высокое смещение).\n", "\n", "Дисперсия: 4.40e-13 — практически нулевая, модель стабильна.\n", "\n", "Вывод: Линейная регрессия не справляется с учетом сложных взаимосвязей между признаками.\n", "\n", "**Градиентный бустинг:**\n", "\n", "MAE: 282.34 — самая низкая ошибка, лучшая модель.\n", "\n", "R²: 0.9791 — объясняет 97.9% дисперсии, хорошо справляется с нелинейными зависимостями.\n", "\n", "Смещение: -696.99 — приемлемое смещение, модель лучше оценивает данные.\n", "\n", "Дисперсия: 4.39 — модель стабильна.\n", "\n", "Вывод: Градиентный бустинг — наиболее точная и сбалансированная модель.\n", "\n", "**Случайные леса:**\n", "\n", "MAE: 292.26 — немного хуже, чем у градиентного бустинга.\n", "\n", "R²: 0.9743 — объясняет 97.4% дисперсии.\n", "\n", "Смещение: -882.53 — модель недооценивает значения сильнее, чем градиентный бустинг.\n", "\n", "Дисперсия: 11.13 — несколько выше, чем у других моделей.\n", "\n", "Вывод: Случайные леса также подходят для задачи, но уступают градиентному бустингу.\n", "\n", "**Градиентный бустинг выбран как лучшая модель из-за низкого MAE и сбалансированных смещения и дисперсии. Случайные леса также могут быть использованы, но с дополнительной настройкой гиперпараметров.**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Задача классификация**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### **Бизнес-цель для задачи регрессии** \n", "\n", "Спрогнозировать категорию огранки бриллианта.\n", "\n", "Это поможет:\n", "\n", "1. Улучшить процессы оценки качества бриллиантов.\n", "2. Упростить классификацию для ювелирных компаний.\n", "3. Повысить точность ценообразования и автоматизировать принятие решений.\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Спрогнозировать категорию категория огранки на основе его характеристик" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Метрики в задаче классификации**\n", "ROC AUC: Измеряет способность модели разделять классы. Чем ближе к 1, тем лучше.\n", "\n", "Точность (Accuracy): Доля правильно классифицированных объектов.\n", "\n", "Смещение: Показывает, насколько среднее значение предсказаний отличается от истинного класса.\n", "\n", "Дисперсия: Описывает изменчивость предсказаний модели при разных обучающих наборах.\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Достижимый уровень качества модели для задачи классификации**\n", "\n", "Целевой уровень: ROC AUC > 0.99, Точность > 0.95.\n", "\n", "Основание: Эти значения обеспечивают высокую точность и надежность при прогнозировании категории огранки бриллианта, минимизируя количество ошибок и повышая доверие к автоматизированной системе оценки качества." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Логистическая регрессия**" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "c:\\D\\semester5\\mii\\aimenv\\Lib\\site-packages\\sklearn\\linear_model\\_logistic.py:1256: FutureWarning: 'multi_class' was deprecated in version 1.5 and will be removed in 1.7. Use OneVsRestClassifier(LogisticRegression(..)) instead. Leave it to its default value to avoid this warning.\n", " warnings.warn(\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Лучшие гиперпараметры: {'classifier__C': 1, 'classifier__penalty': 'l1', 'classifier__solver': 'liblinear'}\n", "ROC AUC у логистической регрессии = 0.9869744388185762\n", "Точность = 0.9338214848456762\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Смещение: 0.9311110918680002\n", "Дисперсия: 0.00184041613203153\n" ] } ], "source": [ "import pandas as pd\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.model_selection import GridSearchCV, train_test_split\n", "from sklearn.metrics import roc_auc_score, confusion_matrix, accuracy_score\n", "import seaborn as sns\n", "import matplotlib.pyplot as plt\n", "from sklearn import metrics\n", "\n", "df = pd.read_csv(\".//static//csv//Diamonds Prices2022.csv\")\n", "\n", "df.columns = df.columns.str.strip()\n", "\n", "# Создание целевой переменной на основе столбца 'price'\n", "bins = [0, 1000, 5000, float('inf')]\n", "labels = ['Low', 'Medium', 'High']\n", "df['price_category'] = pd.cut(df['price'], bins=bins, labels=labels)\n", "\n", "# Преобразуем целевую переменную в категориальный тип\n", "y = pd.Categorical(df['price_category'])\n", "\n", "# Подготовка данных\n", "data = df[['carat', 'cut', 'color', 'clarity', 'depth', 'table', 'x', 'y', 'z', 'price_category']] # Целевая переменная - 'price_category'\n", "X = data.drop('price_category', axis=1) # Признаки\n", "y = data['price_category'] # Целевая переменная\n", "\n", "# Разделение данных на обучающую и тестовую выборки\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n", "\n", "# Преобразование числовых данных\n", "num_imputer = SimpleImputer(strategy=\"median\") # Замена пропусков медианой\n", "num_scaler = StandardScaler() # Стандартизация\n", "preprocessing_num = Pipeline(\n", " [\n", " (\"imputer\", num_imputer),\n", " (\"scaler\", num_scaler),\n", " ]\n", ")\n", "\n", "# Преобразование категориальных данных\n", "cat_imputer = SimpleImputer(strategy=\"most_frequent\") # Заполнение пропусков наиболее частым значением\n", "cat_encoder = OneHotEncoder(handle_unknown='ignore') # Преобразование категориальных признаков в OneHot\n", "preprocessing_cat = Pipeline(\n", " [\n", " (\"imputer\", cat_imputer),\n", " (\"encoder\", cat_encoder),\n", " ]\n", ")\n", "\n", "# Общая предобработка\n", "preprocessing = ColumnTransformer(\n", " [\n", " (\"nums\", preprocessing_num, X.select_dtypes(include=['float64', 'int64']).columns), # Числовые признаки\n", " (\"cats\", preprocessing_cat, X.select_dtypes(include=['object']).columns), # Категориальные признаки\n", " ]\n", ")\n", "\n", "# Конвейер для логистической регрессии\n", "pipeline_logreg = Pipeline([\n", " ('preprocessing', preprocessing),\n", " ('classifier', LogisticRegression(max_iter=1000, multi_class='ovr', solver='liblinear')) # Указание решателя и метода многоклассовой классификации\n", "])\n", "\n", "# Определение сетки гиперпараметров\n", "param_grid = {\n", " 'classifier__C': [0.1, 0.5, 1],\n", " 'classifier__penalty': ['l1', 'l2'],\n", " 'classifier__solver': ['liblinear', 'saga']\n", "}\n", "\n", "# Создание объекта GridSearchCV для поиска лучших гиперпараметров\n", "grid_search = GridSearchCV(pipeline_logreg, param_grid, cv=5, scoring='accuracy', n_jobs=-1)\n", "\n", "# Обучение модели с перебором гиперпараметров\n", "grid_search.fit(X_train, y_train)\n", "\n", "print(\"Лучшие гиперпараметры: \", grid_search.best_params_)\n", "\n", "# Лучшая модель логистической регрессии\n", "best_model = grid_search.best_estimator_\n", "\n", "# Использование и оценка лучшей логистической модели\n", "y_pred_proba = best_model.predict_proba(X_test) # Получаем вероятности для всех классов\n", "\n", "# Для многоклассовой классификации AUC считается для каждого класса\n", "roc_auc = roc_auc_score(y_test, y_pred_proba, multi_class='ovr', average='macro')\n", "print(f'ROC AUC у логистической регрессии = {roc_auc}')\n", "\n", "y_pred = best_model.predict(X_test)\n", "print(f'Точность = {accuracy_score(y_test, y_pred)}')\n", "\n", "# Построение ROC кривой для каждого класса\n", "# Пример для класса \"Medium\"\n", "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba[:, 1], pos_label='Medium')\n", "plt.plot(fpr, tpr)\n", "plt.ylabel('True Positive Rate')\n", "plt.xlabel('False Positive Rate')\n", "plt.title('ROC Curve для класса \"Medium\"')\n", "plt.show()\n", "\n", "# Построение матрицы ошибок\n", "conf_matrix = confusion_matrix(y_test, y_pred)\n", "\n", "# Визуализация матрицы ошибок\n", "plt.figure(figsize=(8, 6))\n", "sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues', \n", " xticklabels=['Предсказанный \"Low\"', 'Предсказанный \"Medium\"', 'Предсказанный \"High\"'], \n", " yticklabels=['Действительно \"Low\"', 'Действительно \"Medium\"', 'Действительно \"High\"'])\n", "plt.title('Confusion Matrix')\n", "plt.ylabel('Actual')\n", "plt.xlabel('Predicted')\n", "plt.show()\n", "\n", "# Оценка смещения и дисперсии\n", "cv_results = grid_search.cv_results_\n", "mean_test_score = cv_results['mean_test_score']\n", "std_test_score = cv_results['std_test_score']\n", "\n", "print(f\"Смещение: {mean_test_score.mean()}\")\n", "print(f\"Дисперсия: {std_test_score.mean()}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Градиентный букинг**" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Лучшие гиперпараметры: {'classifier__learning_rate': 0.1, 'classifier__max_depth': 7, 'classifier__n_estimators': 100, 'classifier__subsample': 0.8}\n", "ROC AUC у градиентного бустинга = 0.99543134281693\n", "Точность = 0.9507832051163222\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Смещение: 0.9380732054788211\n", "Дисперсия: 0.003201824240012704\n" ] } ], "source": [ "import pandas as pd\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.ensemble import GradientBoostingClassifier\n", "from sklearn.model_selection import GridSearchCV, train_test_split\n", "from sklearn.metrics import roc_auc_score, confusion_matrix, accuracy_score\n", "import seaborn as sns\n", "import matplotlib.pyplot as plt\n", "from sklearn import metrics\n", "\n", "df = pd.read_csv(\".//static//csv//Diamonds Prices2022.csv\")\n", "\n", "df.columns = df.columns.str.strip()\n", "\n", "# Создание целевой переменной на основе столбца 'price'\n", "bins = [0, 1000, 5000, float('inf')]\n", "labels = ['Low', 'Medium', 'High']\n", "df['price_category'] = pd.cut(df['price'], bins=bins, labels=labels)\n", "\n", "# Преобразуем целевую переменную в категориальный тип\n", "y = pd.Categorical(df['price_category']) \n", "\n", "# Подготовка данных\n", "data = df[['carat', 'cut', 'color', 'clarity', 'depth', 'table', 'x', 'y', 'z', 'price_category']] # Целевая переменная - 'price_category'\n", "X = data.drop('price_category', axis=1) # Признаки\n", "y = data['price_category'] # Целевая переменная\n", "\n", "# Разделение данных на обучающую и тестовую выборки\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n", "\n", "# Преобразование числовых данных\n", "num_imputer = SimpleImputer(strategy=\"median\") # Замена пропусков медианой\n", "num_scaler = StandardScaler() # Стандартизация\n", "preprocessing_num = Pipeline(\n", " [\n", " (\"imputer\", num_imputer),\n", " (\"scaler\", num_scaler),\n", " ]\n", ")\n", "\n", "# Преобразование категориальных данных\n", "cat_imputer = SimpleImputer(strategy=\"most_frequent\")\n", "cat_encoder = OneHotEncoder(handle_unknown='ignore') \n", "preprocessing_cat = Pipeline(\n", " [\n", " (\"imputer\", cat_imputer),\n", " (\"encoder\", cat_encoder),\n", " ]\n", ")\n", "\n", "# Общая предобработка\n", "preprocessing = ColumnTransformer(\n", " [\n", " (\"nums\", preprocessing_num, X.select_dtypes(include=['float64', 'int64']).columns),\n", " (\"cats\", preprocessing_cat, X.select_dtypes(include=['object']).columns),\n", " ]\n", ")\n", "\n", "# Конвейер для градиентного бустинга\n", "pipeline_gbc = Pipeline([\n", " ('preprocessing', preprocessing),\n", " ('classifier', GradientBoostingClassifier(random_state=42)) # Градиентный бустинг\n", "])\n", "\n", "# Определение сетки гиперпараметров для градиентного бустинга\n", "param_grid = {\n", " 'classifier__n_estimators': [50, 100, 200],\n", " 'classifier__learning_rate': [0.01, 0.1, 0.2],\n", " 'classifier__max_depth': [3, 5, 7],\n", " 'classifier__subsample': [0.8, 1.0]\n", "}\n", "\n", "# Создание объекта GridSearchCV для поиска лучших гиперпараметров\n", "grid_search = GridSearchCV(pipeline_gbc, param_grid, cv=5, scoring='accuracy', n_jobs=-1)\n", "\n", "# Обучение модели с перебором гиперпараметров\n", "grid_search.fit(X_train, y_train)\n", "\n", "print(\"Лучшие гиперпараметры: \", grid_search.best_params_)\n", "\n", "# Лучшая модель градиентного бустинга\n", "best_model = grid_search.best_estimator_\n", "\n", "# Использование и оценка лучшей модели\n", "y_pred_proba = best_model.predict_proba(X_test) # Получаем вероятности для всех классов\n", "\n", "# Для многоклассовой классификации AUC считается для каждого класса\n", "roc_auc = roc_auc_score(y_test, y_pred_proba, multi_class='ovr', average='macro')\n", "print(f'ROC AUC у градиентного бустинга = {roc_auc}')\n", "\n", "y_pred = best_model.predict(X_test)\n", "print(f'Точность = {accuracy_score(y_test, y_pred)}')\n", "\n", "# Построение ROC кривой для каждого класса\n", "# Пример для класса \"Medium\"\n", "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba[:, 1], pos_label='Medium')\n", "plt.plot(fpr, tpr)\n", "plt.ylabel('True Positive Rate')\n", "plt.xlabel('False Positive Rate')\n", "plt.title('ROC Curve для класса \"Medium\"')\n", "plt.show()\n", "\n", "# Построение матрицы ошибок\n", "conf_matrix = confusion_matrix(y_test, y_pred)\n", "\n", "# Визуализация матрицы ошибок\n", "plt.figure(figsize=(8, 6))\n", "sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues', \n", " xticklabels=['Предсказанный \"Low\"', 'Предсказанный \"Medium\"', 'Предсказанный \"High\"'], \n", " yticklabels=['Действительно \"Low\"', 'Действительно \"Medium\"', 'Действительно \"High\"'])\n", "plt.title('Confusion Matrix')\n", "plt.ylabel('Actual')\n", "plt.xlabel('Predicted')\n", "plt.show()\n", "\n", "# Оценка смещения и дисперсии\n", "cv_results = grid_search.cv_results_\n", "mean_test_score = cv_results['mean_test_score']\n", "std_test_score = cv_results['std_test_score']\n", "\n", "print(f\"Смещение: {mean_test_score.mean()}\")\n", "print(f\"Дисперсия: {std_test_score.mean()}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Случайный лес**" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Лучшие гиперпараметры: {'classifier__n_estimators': 100, 'classifier__min_samples_split': 2, 'classifier__min_samples_leaf': 1, 'classifier__max_features': 'sqrt', 'classifier__max_depth': 20}\n", "ROC AUC для случайного леса = 0.9946705103108209\n", "Точность = 0.9466122902956715\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Смещение: 0.942448004061997\n", "Дисперсия: 0.003142805279982323\n" ] } ], "source": [ "import pandas as pd\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.model_selection import RandomizedSearchCV, train_test_split\n", "from sklearn.metrics import roc_auc_score, confusion_matrix, accuracy_score\n", "import seaborn as sns\n", "import matplotlib.pyplot as plt\n", "from sklearn import metrics\n", "\n", "df = pd.read_csv(\".//static//csv//Diamonds Prices2022.csv\")\n", "\n", "# Убираем пробелы в именах столбцов, если есть\n", "df.columns = df.columns.str.strip()\n", "\n", "# Создание целевой переменной на основе столбца 'price'\n", "bins = [0, 1000, 5000, float('inf')]\n", "labels = ['Low', 'Medium', 'High']\n", "df['price_category'] = pd.cut(df['price'], bins=bins, labels=labels)\n", "\n", "# Преобразуем целевую переменную в категориальный тип\n", "y = pd.Categorical(df['price_category'])\n", "\n", "# Подготовка данных\n", "data = df[['carat', 'cut', 'color', 'clarity', 'depth', 'table', 'x', 'y', 'z', 'price_category']] # Целевая переменная - 'price_category'\n", "X = data.drop('price_category', axis=1) # Признаки\n", "y = data['price_category'] # Целевая переменная\n", "\n", "# Разделение данных на обучающую и тестовую выборки\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n", "\n", "# Преобразование числовых данных\n", "num_imputer = SimpleImputer(strategy=\"median\") # Замена пропусков медианой\n", "num_scaler = StandardScaler() # Стандартизация\n", "preprocessing_num = Pipeline(\n", " [\n", " (\"imputer\", num_imputer),\n", " (\"scaler\", num_scaler),\n", " ]\n", ")\n", "\n", "# Преобразование категориальных данных\n", "cat_imputer = SimpleImputer(strategy=\"most_frequent\") # Заполнение пропусков наиболее частым значением\n", "cat_encoder = OneHotEncoder(handle_unknown='ignore') # Преобразование категориальных признаков в OneHot\n", "preprocessing_cat = Pipeline(\n", " [\n", " (\"imputer\", cat_imputer),\n", " (\"encoder\", cat_encoder),\n", " ]\n", ")\n", "\n", "# Общая предобработка\n", "preprocessing = ColumnTransformer(\n", " [\n", " (\"nums\", preprocessing_num, X.select_dtypes(include=['float64', 'int64']).columns), # Числовые признаки\n", " (\"cats\", preprocessing_cat, X.select_dtypes(include=['object']).columns), # Категориальные признаки\n", " ]\n", ")\n", "\n", "# Конвейер для случайного леса\n", "pipeline_rf = Pipeline([\n", " ('preprocessing', preprocessing),\n", " ('classifier', RandomForestClassifier(random_state=42)) # Модель случайного леса\n", "])\n", "\n", "# Определение сетки гиперпараметров\n", "param_grid = {\n", " 'classifier__n_estimators': [50, 100, 150], # Количество деревьев\n", " 'classifier__max_depth': [10, 20, None], # Глубина деревьев\n", " 'classifier__min_samples_split': [2, 5, 10], # Минимальное количество образцов для разделения\n", " 'classifier__min_samples_leaf': [1, 2, 4], # Минимальное количество образцов в листьях\n", " 'classifier__max_features': ['sqrt', 'log2'] # Количество признаков для каждого дерева\n", "}\n", "\n", "# Создание объекта RandomizedSearchCV для поиска лучших гиперпараметров\n", "random_search = RandomizedSearchCV(pipeline_rf, param_grid, n_iter=10, cv=5, scoring='accuracy', n_jobs=-1, random_state=42)\n", "\n", "# Обучение модели с перебором гиперпараметров\n", "random_search.fit(X_train, y_train)\n", "\n", "print(\"Лучшие гиперпараметры: \", random_search.best_params_)\n", "\n", "# Лучшая модель случайного леса\n", "best_model = random_search.best_estimator_\n", "\n", "# Использование и оценка лучшей модели случайного леса\n", "y_pred_proba = best_model.predict_proba(X_test) # Получаем вероятности для всех классов\n", "\n", "# Для многоклассовой классификации используем roc_auc_score с multi_class='ovr'\n", "print(f'ROC AUC для случайного леса = {roc_auc_score(y_test, y_pred_proba, multi_class=\"ovr\")}')\n", "\n", "y_pred = best_model.predict(X_test)\n", "print(f'Точность = {accuracy_score(y_test, y_pred)}')\n", "\n", "# Построение ROC кривой для многоклассовой задачи\n", "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba[:, 2], pos_label='High') # Указываем класс для которого строим ROC\n", "plt.plot(fpr, tpr)\n", "plt.ylabel('True Positive Rate')\n", "plt.xlabel('False Positive Rate')\n", "plt.title('ROC Curve')\n", "plt.show()\n", "\n", "# Построение матрицы ошибок\n", "conf_matrix = confusion_matrix(y_test, y_pred)\n", "\n", "# Визуализация матрицы ошибок\n", "plt.figure(figsize=(8, 6))\n", "sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues', \n", " xticklabels=['Предсказанный \"Low\"', 'Предсказанный \"Medium\"', 'Предсказанный \"High\"'], \n", " yticklabels=['Действительно \"Low\"', 'Действительно \"Medium\"', 'Действительно \"High\"'])\n", "plt.title('Confusion Matrix')\n", "plt.ylabel('Actual')\n", "plt.xlabel('Predicted')\n", "plt.show()\n", "\n", "\n", "# Оценка смещения и дисперсии\n", "cv_results = random_search.cv_results_\n", "mean_test_score = cv_results['mean_test_score']\n", "std_test_score = cv_results['std_test_score']\n", "\n", "print(f\"Смещение: {mean_test_score.mean()}\")\n", "print(f\"Дисперсия: {std_test_score.mean()}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### **Вывод:**\n", "\n", "Градиентный бустинг показал наивысший ROC AUC (0.995) и точность (0.951), но имеет высокую дисперсию (0.0032). Это указывает на более сложные зависимости, которые модель способна уловить.\n", "\n", "Случайный лес близок по качеству к градиентному бустингу (ROC AUC = 0.995, точность = 0.947), при этом демонстрирует меньшую дисперсию. Это делает его более стабильным при изменении данных.\n", "\n", "Логистическая регрессия уступает обоим ансамблевым методам, но имеет минимальную дисперсию (0.0018), что делает её менее подверженной переобучению." ] } ], "metadata": { "kernelspec": { "display_name": "aimenv", "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": "3.12.6" } }, "nbformat": 4, "nbformat_minor": 2 }