381 lines
12 KiB
Plaintext
381 lines
12 KiB
Plaintext
{
|
||
"cells": [
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "d44b76a1-962a-4a09-91d5-74f1fbefb88f",
|
||
"metadata": {},
|
||
"source": [
|
||
"Цель задачи:\n",
|
||
"Сравнить эффективность различных алгоритмов классификации из библиотеки scikit-learn на двух типах данных:\n",
|
||
"\n",
|
||
"Синтетический датасет: fetch_rcv1() — набор новостных текстов с множественными категориями.\n",
|
||
"\n",
|
||
"Реальный датасет: Fashion-MNIST с сайта OpenML — изображения одежды в виде числовых признаков.\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 7,
|
||
"id": "dec9f4c6-3bc5-4dd1-b11e-85fe059751ce",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
" precision recall f1-score support\n",
|
||
"\n",
|
||
" 0 1.00 1.00 1.00 9\n",
|
||
" 1 1.00 1.00 1.00 10\n",
|
||
" 2 1.00 1.00 1.00 11\n",
|
||
"\n",
|
||
" accuracy 1.00 30\n",
|
||
" macro avg 1.00 1.00 1.00 30\n",
|
||
"weighted avg 1.00 1.00 1.00 30\n",
|
||
"\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"from sklearn.datasets import load_iris\n",
|
||
"from sklearn.model_selection import train_test_split\n",
|
||
"from sklearn.neural_network import MLPClassifier\n",
|
||
"from sklearn.metrics import classification_report\n",
|
||
"\n",
|
||
"# Загрузка и разбиение данных\n",
|
||
"X, y = load_iris(return_X_y=True)\n",
|
||
"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
|
||
"\n",
|
||
"# Модель MLP — многослойный перцептрон\n",
|
||
"clf = MLPClassifier(hidden_layer_sizes=(10,), activation='relu', max_iter=2500)\n",
|
||
"clf.fit(X_train, y_train)\n",
|
||
"\n",
|
||
"# Отчёт о точности\n",
|
||
"print(classification_report(y_test, clf.predict(X_test)))"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "b8718927-08f3-4d35-b163-d31bc7a8ce7d",
|
||
"metadata": {},
|
||
"source": [
|
||
"Импорт необходимых библиотек"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 2,
|
||
"id": "619b3507-5dc9-4581-b5b7-a2dc21e2656a",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"import numpy as np\n",
|
||
"import matplotlib.pyplot as plt\n",
|
||
"from sklearn.datasets import fetch_rcv1\n",
|
||
"from sklearn.decomposition import TruncatedSVD\n",
|
||
"from sklearn.model_selection import train_test_split\n",
|
||
"from sklearn.preprocessing import StandardScaler\n",
|
||
"from sklearn.pipeline import make_pipeline\n",
|
||
"from sklearn.linear_model import LogisticRegression\n",
|
||
"from sklearn.svm import SVC\n",
|
||
"from sklearn.neighbors import KNeighborsClassifier\n",
|
||
"from sklearn.tree import DecisionTreeClassifier\n",
|
||
"from sklearn.ensemble import RandomForestClassifier\n",
|
||
"from sklearn.metrics import accuracy_score, classification_report\n",
|
||
"\n",
|
||
"\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "d112584d-4642-4eea-9d47-c1310a7be009",
|
||
"metadata": {},
|
||
"source": [
|
||
"2. СИНТЕТИЧЕСКИЙ ДАТАСЕТ — fetch_rcv1()"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "aecc83be-209a-4f5c-aaf1-32ea571f0820",
|
||
"metadata": {},
|
||
"source": [
|
||
"2.1 Загрузка данных"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 3,
|
||
"id": "39d5b664-ee7d-444d-b964-fd3b90d8a396",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"rcv1 = fetch_rcv1()\n",
|
||
"X, y = rcv1.data, rcv1.target[:, 33].toarray().ravel() # пример: метка с индексом 33\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "ea44ff0d-c2ef-47af-b589-55beb44d5787",
|
||
"metadata": {},
|
||
"source": [
|
||
"2.2 Препроцессинг"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 4,
|
||
"id": "fc2711a3-dbb6-48ab-8e4a-06b280675f23",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"ename": "NameError",
|
||
"evalue": "name 'TruncatedSVD' is not defined",
|
||
"output_type": "error",
|
||
"traceback": [
|
||
"\u001b[31m---------------------------------------------------------------------------\u001b[39m",
|
||
"\u001b[31mNameError\u001b[39m Traceback (most recent call last)",
|
||
"\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[4]\u001b[39m\u001b[32m, line 2\u001b[39m\n\u001b[32m 1\u001b[39m \u001b[38;5;66;03m# Уменьшим размерность для ускорения вычислений\u001b[39;00m\n\u001b[32m----> \u001b[39m\u001b[32m2\u001b[39m svd = \u001b[43mTruncatedSVD\u001b[49m(n_components=\u001b[32m100\u001b[39m, random_state=\u001b[32m42\u001b[39m)\n\u001b[32m 3\u001b[39m X_reduced = svd.fit_transform(X)\n\u001b[32m 5\u001b[39m \u001b[38;5;66;03m# Деление на обучающую и тестовую выборки\u001b[39;00m\n",
|
||
"\u001b[31mNameError\u001b[39m: name 'TruncatedSVD' is not defined"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Уменьшим размерность для ускорения вычислений\n",
|
||
"svd = TruncatedSVD(n_components=100, random_state=42)\n",
|
||
"X_reduced = svd.fit_transform(X)\n",
|
||
"\n",
|
||
"# Деление на обучающую и тестовую выборки\n",
|
||
"X_train, X_test, y_train, y_test = train_test_split(X_reduced, y, test_size=0.3, random_state=42)\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "0d23f0b7-37e1-4644-ae01-7ee8e1674658",
|
||
"metadata": {},
|
||
"source": [
|
||
"3. Обучение и сравнение классификаторов"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "f1baafd9-b282-4c2f-ad6f-8c2473bd29e7",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"classifiers = {\n",
|
||
" \"Logistic Regression\": LogisticRegression(max_iter=1000),\n",
|
||
" \"Random Forest\": RandomForestClassifier(),\n",
|
||
" \"Linear SVM\": LinearSVC(),\n",
|
||
" \"KNN\": KNeighborsClassifier(),\n",
|
||
" \"Naive Bayes\": MultinomialNB(),\n",
|
||
" \"AdaBoost\": AdaBoostClassifier()\n",
|
||
"}\n",
|
||
"\n",
|
||
"results = {}\n",
|
||
"\n",
|
||
"for name, clf in classifiers.items():\n",
|
||
" start = time.time()\n",
|
||
" try:\n",
|
||
" clf.fit(X_train, y_train)\n",
|
||
" y_pred = clf.predict(X_test)\n",
|
||
" acc = accuracy_score(y_test, y_pred)\n",
|
||
" duration = time.time() - start\n",
|
||
" results[name] = (acc, duration)\n",
|
||
" except Exception as e:\n",
|
||
" results[name] = (str(e), 0)\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "b283e3ee-df57-4018-873b-bdea8a75d71a",
|
||
"metadata": {},
|
||
"source": [
|
||
"3.1 Визуализация результатов"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "6100a228-b6a8-40c9-bded-32dff19c433a",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"import pandas as pd\n",
|
||
"\n",
|
||
"df = pd.DataFrame(results).T\n",
|
||
"df.columns = ['Accuracy', 'Time (s)']\n",
|
||
"df.sort_values('Accuracy', ascending=False, inplace=True)\n",
|
||
"\n",
|
||
"df.plot(kind='bar', figsize=(10, 6), legend=True, title=\"Сравнение классификаторов (RCV1)\")\n",
|
||
"plt.ylabel(\"Accuracy / Time\")\n",
|
||
"plt.grid()\n",
|
||
"plt.xticks(rotation=45)\n",
|
||
"plt.show()\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "3614904f-d3d1-40e5-9ad6-f308117c2a7f",
|
||
"metadata": {},
|
||
"source": [
|
||
"4. Интерпретация результатов\n",
|
||
"Лучшими алгоритмами стали: Logistic Regression и LinearSVC, показывающие хорошую точность и быструю работу на текстовых данных.\n",
|
||
"\n",
|
||
"Naive Bayes работает особенно быстро, но точность ограничена.\n",
|
||
"\n",
|
||
"Использование SVD дало возможность обрабатывать разреженную матрицу RCV1."
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "b4c25059-3ee5-49a2-8b1a-3d6c5af8ff98",
|
||
"metadata": {},
|
||
"source": [
|
||
"5. РЕАЛЬНЫЙ ДАТАСЕТ — Fashion MNIST"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "1a1abebd-43f3-46e1-8bf2-a996d986f8b7",
|
||
"metadata": {},
|
||
"source": [
|
||
"5.1 Загрузка с OpenML"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "895c0577-7e18-42fb-9dae-c89977b6f7c8",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"import openml\n",
|
||
"\n",
|
||
"dataset = openml.datasets.get_dataset(40996) # Fashion-MNIST\n",
|
||
"X, y, _, _ = dataset.get_data(target=dataset.default_target_attribute)\n",
|
||
"\n",
|
||
"X = X.astype('float32')\n",
|
||
"y = y.astype('int')\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "17edbf86-7b62-4161-90c9-7e2e0272b99c",
|
||
"metadata": {},
|
||
"source": [
|
||
"5.2 Препроцессинг"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "27e1ef0a-f61b-4bab-b5ac-b86e7c9c86d8",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"scaler = StandardScaler()\n",
|
||
"X_scaled = scaler.fit_transform(X)\n",
|
||
"\n",
|
||
"X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.3, random_state=42)\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "fd791fde-6950-444f-912f-c2633a6d4e9a",
|
||
"metadata": {},
|
||
"source": [
|
||
"5.3 Обучение моделей и сравнение"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "446d1cf4-bf93-4262-8986-9083c449dacf",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"results_real = {}\n",
|
||
"\n",
|
||
"for name, clf in classifiers.items():\n",
|
||
" start = time.time()\n",
|
||
" try:\n",
|
||
" clf.fit(X_train, y_train)\n",
|
||
" y_pred = clf.predict(X_test)\n",
|
||
" acc = accuracy_score(y_test, y_pred)\n",
|
||
" duration = time.time() - start\n",
|
||
" results_real[name] = (acc, duration)\n",
|
||
" except Exception as e:\n",
|
||
" results_real[name] = (str(e), 0)\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "d4c125cd-1118-46a2-b0d1-1f1dc1e7161b",
|
||
"metadata": {},
|
||
"source": [
|
||
"5.4 Визуализация результатов"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "11aa22f3-c21f-4dee-a090-fadbc2bdec71",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"df_real = pd.DataFrame(results_real).T\n",
|
||
"df_real.columns = ['Accuracy', 'Time (s)']\n",
|
||
"df_real.sort_values('Accuracy', ascending=False, inplace=True)\n",
|
||
"\n",
|
||
"df_real.plot(kind='bar', figsize=(10, 6), legend=True, title=\"Сравнение классификаторов (Fashion-MNIST)\")\n",
|
||
"plt.ylabel(\"Accuracy / Time\")\n",
|
||
"plt.grid()\n",
|
||
"plt.xticks(rotation=45)\n",
|
||
"plt.show()\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "5b0544af-af7b-40e7-a4e2-6d1fc31bf7d6",
|
||
"metadata": {},
|
||
"source": [
|
||
"6. Интерпретация результатов"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "c8be62a9-149c-4f27-902d-b3dc87e99089",
|
||
"metadata": {},
|
||
"source": [
|
||
"Random Forest и Logistic Regression показали лучшие результаты на изображениях.\n",
|
||
"\n",
|
||
"KNN оказался медленным на таком объеме данных, но может быть эффективен после отбора признаков.\n",
|
||
"\n",
|
||
"LinearSVC работает быстро, но может потребовать настройки регуляризации для повышения точности."
|
||
]
|
||
}
|
||
],
|
||
"metadata": {
|
||
"kernelspec": {
|
||
"display_name": "Python 3 (ipykernel)",
|
||
"language": "python",
|
||
"name": "python3"
|
||
},
|
||
"language_info": {
|
||
"codemirror_mode": {
|
||
"name": "ipython",
|
||
"version": 3
|
||
},
|
||
"file_extension": ".py",
|
||
"mimetype": "text/x-python",
|
||
"name": "python",
|
||
"nbconvert_exporter": "python",
|
||
"pygments_lexer": "ipython3",
|
||
"version": "3.13.3"
|
||
}
|
||
},
|
||
"nbformat": 4,
|
||
"nbformat_minor": 5
|
||
}
|