{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# 机器学习分类预测分析 - 完整工作流\n", "\n", "## 项目背景\n", "本Notebook演示了完整的机器学习数据分析流程，使用二分类数据集预测目标变量Y。\n", "\n", "## 分析流程\n", "1. 数据加载与概览\n", "2. 数据清洗与预处理\n", "3. 探索性数据分析(EDA)\n", "4. 特征工程\n", "5. 模型训练与评估\n", "6. 结果可视化与解释" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:32.858833Z", "iopub.status.busy": "2026-03-20T02:20:32.858429Z", "iopub.status.idle": "2026-03-20T02:20:36.034989Z", "shell.execute_reply": "2026-03-20T02:20:36.034301Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "✅ 库导入成功！\n" ] } ], "source": [ "# 导入必要的库\n", "import pandas as pd\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "from matplotlib import font_manager\n", "import warnings\n", "warnings.filterwarnings('ignore')\n", "\n", "# 设置中文字体\n", "font_path = '/System/Library/Fonts/STHeiti Medium.ttc'\n", "plt.rcParams['font.family'] = font_manager.FontProperties(fname=font_path).get_name()\n", "plt.rcParams['axes.unicode_minus'] = False\n", "\n", "# 设置绘图样式\n", "plt.style.use('seaborn-v0_8')\n", "sns.set_palette(\"husl\")\n", "\n", "# 导入机器学习库\n", "from sklearn.model_selection import train_test_split, cross_val_score\n", "from sklearn.preprocessing import LabelEncoder, StandardScaler\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.tree import DecisionTreeClassifier\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.metrics import (\n", " accuracy_score, precision_score, recall_score, f1_score, \n", " roc_auc_score, confusion_matrix, classification_report, \n", " roc_curve\n", ")\n", "\n", "print(\"✅ 库导入成功！\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 1. 数据加载与概览" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:36.073740Z", "iopub.status.busy": "2026-03-20T02:20:36.073303Z", "iopub.status.idle": "2026-03-20T02:20:36.107847Z", "shell.execute_reply": "2026-03-20T02:20:36.107294Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "============================================================\n", "数据概览\n", "============================================================\n", "\n", "数据形状: 4050 行 × 8 列\n", "\n", "列名: ['p1', 'p2', 'gender', 'family', 'realty_ind', 'age', 'p3_f', 'Y']\n", "\n", "数据类型:\n", "p1 int64\n", "p2 int64\n", "gender object\n", "family object\n", "realty_ind int64\n", "age int64\n", "p3_f int64\n", "Y int64\n", "dtype: object\n" ] } ], "source": [ "# 加载数据\n", "df = pd.read_csv('3.3.预测型分析 - 机器学习.csv', encoding='utf-8-sig')\n", "\n", "# 基础信息\n", "print(\"=\"*60)\n", "print(\"数据概览\")\n", "print(\"=\"*60)\n", "print(f\"\\n数据形状: {df.shape[0]} 行 × {df.shape[1]} 列\")\n", "print(f\"\\n列名: {list(df.columns)}\")\n", "print(f\"\\n数据类型:\")\n", "print(df.dtypes)" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:36.110196Z", "iopub.status.busy": "2026-03-20T02:20:36.109981Z", "iopub.status.idle": "2026-03-20T02:20:36.124621Z", "shell.execute_reply": "2026-03-20T02:20:36.124157Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "数据预览（前10行）:\n" ] }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	p1	p2	gender	family	realty_ind	age	p3_f	Y
0	0	1	F	Married	1	57	6	0
1	1	1	M	Married	1	42	18	1
2	0	1	F	Married	1	65	12	0
3	0	1	M	Civil marriage	1	56	6	1
4	1	1	M	Single / not married	1	39	18	1
5	0	1	F	Married	1	47	6	0
6	0	1	M	Single / not married	1	41	0	1
7	0	1	F	Married	1	31	18	1
8	1	1	F	Married	1	55	24	1
9	0	1	F	Civil marriage	1	66	6	0

\n", "

" ], "text/plain": [ " p1 p2 gender family realty_ind age p3_f Y\n", "0 0 1 F Married 1 57 6 0\n", "1 1 1 M Married 1 42 18 1\n", "2 0 1 F Married 1 65 12 0\n", "3 0 1 M Civil marriage 1 56 6 1\n", "4 1 1 M Single / not married 1 39 18 1\n", "5 0 1 F Married 1 47 6 0\n", "6 0 1 M Single / not married 1 41 0 1\n", "7 0 1 F Married 1 31 18 1\n", "8 1 1 F Married 1 55 24 1\n", "9 0 1 F Civil marriage 1 66 6 0" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 查看前几行数据\n", "print(\"\\n数据预览（前10行）:\")\n", "df.head(10)" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:36.127161Z", "iopub.status.busy": "2026-03-20T02:20:36.126916Z", "iopub.status.idle": "2026-03-20T02:20:36.148574Z", "shell.execute_reply": "2026-03-20T02:20:36.148144Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "============================================================\n", "数据质量检查\n", "============================================================\n", "\n", "缺失值统计:\n", "无缺失值\n", "\n", "重复行数: 1732\n", "\n", "目标变量Y分布:\n", "Y\n", "0 2762\n", "1 1288\n", "Name: count, dtype: int64\n", "\n", "目标变量比例:\n", "Y\n", "0 0.682\n", "1 0.318\n", "Name: proportion, dtype: float64\n" ] } ], "source": [ "# 数据质量检查\n", "print(\"\\n\" + \"=\"*60)\n", "print(\"数据质量检查\")\n", "print(\"=\"*60)\n", "\n", "# 缺失值统计\n", "missing = df.isnull().sum()\n", "print(f\"\\n缺失值统计:\")\n", "print(missing[missing > 0] if missing.sum() > 0 else \"无缺失值\")\n", "\n", "# 重复值检查\n", "duplicates = df.duplicated().sum()\n", "print(f\"\\n重复行数: {duplicates}\")\n", "\n", "# 目标变量分布\n", "print(f\"\\n目标变量Y分布:\")\n", "print(df['Y'].value_counts())\n", "print(f\"\\n目标变量比例:\")\n", "print(df['Y'].value_counts(normalize=True).round(4))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 2. 探索性数据分析(EDA)" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:36.150977Z", "iopub.status.busy": "2026-03-20T02:20:36.150750Z", "iopub.status.idle": "2026-03-20T02:20:36.170423Z", "shell.execute_reply": "2026-03-20T02:20:36.169907Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "数值型变量描述统计:\n" ] }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	p1	p2	realty_ind	age	p3_f	Y
count	4050.000000	4050.000000	4050.000000	4050.000000	4050.000000	4050.000000
mean	0.114074	0.971852	0.688889	43.626173	7.506173	0.318025
std	0.317940	0.165416	0.463005	11.396685	6.408972	0.465767
min	0.000000	0.000000	0.000000	22.000000	0.000000	0.000000
25%	0.000000	1.000000	0.000000	34.000000	4.000000	0.000000
50%	0.000000	1.000000	1.000000	43.000000	6.000000	0.000000
75%	0.000000	1.000000	1.000000	53.000000	12.000000	1.000000
max	1.000000	1.000000	1.000000	68.000000	30.000000	1.000000

\n", "

" ], "text/plain": [ " p1 p2 realty_ind age p3_f \\\n", "count 4050.000000 4050.000000 4050.000000 4050.000000 4050.000000 \n", "mean 0.114074 0.971852 0.688889 43.626173 7.506173 \n", "std 0.317940 0.165416 0.463005 11.396685 6.408972 \n", "min 0.000000 0.000000 0.000000 22.000000 0.000000 \n", "25% 0.000000 1.000000 0.000000 34.000000 4.000000 \n", "50% 0.000000 1.000000 1.000000 43.000000 6.000000 \n", "75% 0.000000 1.000000 1.000000 53.000000 12.000000 \n", "max 1.000000 1.000000 1.000000 68.000000 30.000000 \n", "\n", " Y \n", "count 4050.000000 \n", "mean 0.318025 \n", "std 0.465767 \n", "min 0.000000 \n", "25% 0.000000 \n", "50% 0.000000 \n", "75% 1.000000 \n", "max 1.000000 " ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 数值型变量描述统计\n", "print(\"数值型变量描述统计:\")\n", "df.describe()" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:36.172806Z", "iopub.status.busy": "2026-03-20T02:20:36.172580Z", "iopub.status.idle": "2026-03-20T02:20:36.178917Z", "shell.execute_reply": "2026-03-20T02:20:36.178516Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "类别型变量统计:\n", "\n", "gender:\n", "gender\n", "F 2663\n", "M 1387\n", "Name: count, dtype: int64\n", "\n", "family:\n", "family\n", "Married 2750\n", "Single / not married 544\n", "Civil marriage 355\n", "Separated 237\n", "Widow 164\n", "Name: count, dtype: int64\n" ] } ], "source": [ "# 类别型变量统计\n", "print(\"\\n类别型变量统计:\")\n", "categorical_cols = ['gender', 'family']\n", "for col in categorical_cols:\n", " print(f\"\\n{col}:\")\n", " print(df[col].value_counts().head(10))" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:36.181283Z", "iopub.status.busy": "2026-03-20T02:20:36.181040Z", "iopub.status.idle": "2026-03-20T02:20:38.001603Z", "shell.execute_reply": "2026-03-20T02:20:38.001058Z" } }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "

" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "✅ EDA可视化已生成\n" ] } ], "source": [ "# 创建综合可视化图\n", "fig, axes = plt.subplots(2, 3, figsize=(18, 12))\n", "\n", "# 1. 目标变量分布\n", "ax1 = axes[0, 0]\n", "y_counts = df['Y'].value_counts()\n", "colors = ['#FF6B6B', '#4ECDC4']\n", "ax1.pie(y_counts, labels=['Negative (0)', 'Positive (1)'], autopct='%1.1f%%', \n", " colors=colors, startangle=90)\n", "ax1.set_title('目标变量Y分布', fontsize=14, fontweight='bold')\n", "\n", "# 2. 性别分布与目标变量关系\n", "ax2 = axes[0, 1]\n", "gender_y = pd.crosstab(df['gender'], df['Y'], normalize='index') * 100\n", "gender_y.plot(kind='bar', ax=ax2, color=colors)\n", "ax2.set_title('性别 vs 目标变量（%）', fontsize=14, fontweight='bold')\n", "ax2.set_xlabel('性别')\n", "ax2.set_ylabel('百分比')\n", "ax2.legend(['Y=0', 'Y=1'])\n", "ax2.tick_params(axis='x', rotation=0)\n", "\n", "# 3. 家庭状态分布\n", "ax3 = axes[0, 2]\n", "family_counts = df['family'].value_counts()\n", "ax3.barh(family_counts.index, family_counts.values, color='skyblue')\n", "ax3.set_title('家庭状态分布', fontsize=14, fontweight='bold')\n", "ax3.set_xlabel('数量')\n", "\n", "# 4. 年龄分布\n", "ax4 = axes[1, 0]\n", "ax4.hist(df['age'], bins=30, color='coral', edgecolor='black', alpha=0.7)\n", "ax4.axvline(df['age'].mean(), color='red', linestyle='--', linewidth=2, label=f'均值: {df[\"age\"].mean():.1f}')\n", "ax4.axvline(df['age'].median(), color='blue', linestyle='--', linewidth=2, label=f'中位数: {df[\"age\"].median():.1f}')\n", "ax4.set_title('年龄分布', fontsize=14, fontweight='bold')\n", "ax4.set_xlabel('年龄')\n", "ax4.set_ylabel('频数')\n", "ax4.legend()\n", "\n", "# 5. 年龄 vs 目标变量\n", "ax5 = axes[1, 1]\n", "df.boxplot(column='age', by='Y', ax=ax5)\n", "ax5.set_title('年龄分布 by 目标变量', fontsize=14, fontweight='bold')\n", "ax5.set_xlabel('目标变量 Y')\n", "ax5.set_ylabel('年龄')\n", "plt.suptitle('')\n", "\n", "# 6. 房产指标 vs 目标变量\n", "ax6 = axes[1, 2]\n", "realty_y = pd.crosstab(df['realty_ind'], df['Y'], normalize='index') * 100\n", "realty_y.plot(kind='bar', ax=ax6, color=colors)\n", "ax6.set_title('房产指标 vs 目标变量（%）', fontsize=14, fontweight='bold')\n", "ax6.set_xlabel('房产指标 (0=无, 1=有)')\n", "ax6.set_ylabel('百分比')\n", "ax6.legend(['Y=0', 'Y=1'])\n", "ax6.tick_params(axis='x', rotation=0)\n", "\n", "plt.tight_layout()\n", "plt.savefig('eda_overview.png', dpi=150, bbox_inches='tight')\n", "plt.show()\n", "\n", "print(\"✅ EDA可视化已生成\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 3. 特征工程与数据预处理" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:38.005238Z", "iopub.status.busy": "2026-03-20T02:20:38.004846Z", "iopub.status.idle": "2026-03-20T02:20:38.028735Z", "shell.execute_reply": "2026-03-20T02:20:38.028086Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "============================================================\n", "特征工程\n", "============================================================\n", "\n", "标签编码映射:\n", "gender: {'F': 0, 'M': 1}\n", "family: {'Civil marriage': 0, 'Married': 1, 'Separated': 2, 'Single / not married': 3, 'Widow': 4}\n", "\n", "新增特征:\n", "- gender_encoded: 性别编码\n", "- family_encoded: 家庭状态编码\n", "- age_group: 年龄分组\n", "- has_realty_and_family: 有房产且已婚\n", "- p_product: p1和p2的乘积\n", "\n", "处理后数据形状: (4050, 13)\n" ] }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	p1	p2	gender	family	realty_ind	age	p3_f	Y	gender_encoded	family_encoded	age_group	has_realty_and_family	p_product
0	0	1	F	Married	1	57	6	0	0	1	50-60	1	0
1	1	1	M	Married	1	42	18	1	1	1	40-50	1	1
2	0	1	F	Married	1	65	12	0	0	1	60+	1	0
3	0	1	M	Civil marriage	1	56	6	1	1	0	50-60	0	0
4	1	1	M	Single / not married	1	39	18	1	1	3	30-40	0	1

\n", "

" ], "text/plain": [ " p1 p2 gender family realty_ind age p3_f Y \\\n", "0 0 1 F Married 1 57 6 0 \n", "1 1 1 M Married 1 42 18 1 \n", "2 0 1 F Married 1 65 12 0 \n", "3 0 1 M Civil marriage 1 56 6 1 \n", "4 1 1 M Single / not married 1 39 18 1 \n", "\n", " gender_encoded family_encoded age_group has_realty_and_family p_product \n", "0 0 1 50-60 1 0 \n", "1 1 1 40-50 1 1 \n", "2 0 1 60+ 1 0 \n", "3 1 0 50-60 0 0 \n", "4 1 3 30-40 0 1 " ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 创建数据副本进行处理\n", "df_processed = df.copy()\n", "\n", "print(\"=\"*60)\n", "print(\"特征工程\")\n", "print(\"=\"*60)\n", "\n", "# 1. 标签编码\n", "le_gender = LabelEncoder()\n", "le_family = LabelEncoder()\n", "\n", "df_processed['gender_encoded'] = le_gender.fit_transform(df_processed['gender'])\n", "df_processed['family_encoded'] = le_family.fit_transform(df_processed['family'])\n", "\n", "print(f\"\\n标签编码映射:\")\n", "print(f\"gender: {dict(zip(le_gender.classes_, le_gender.transform(le_gender.classes_)))}\")\n", "print(f\"family: {dict(zip(le_family.classes_, le_family.transform(le_family.classes_)))}\")\n", "\n", "# 2. 年龄分段\n", "df_processed['age_group'] = pd.cut(df_processed['age'], \n", " bins=[0, 30, 40, 50, 60, 100], \n", " labels=['<30', '30-40', '40-50', '50-60', '60+'])\n", "\n", "# 3. 创建交互特征\n", "df_processed['has_realty_and_family'] = df_processed['realty_ind'] * (df_processed['family'] == 'Married').astype(int)\n", "df_processed['p_product'] = df_processed['p1'] * df_processed['p2']\n", "\n", "print(f\"\\n新增特征:\")\n", "print(\"- gender_encoded: 性别编码\")\n", "print(\"- family_encoded: 家庭状态编码\")\n", "print(\"- age_group: 年龄分组\")\n", "print(\"- has_realty_and_family: 有房产且已婚\")\n", "print(\"- p_product: p1和p2的乘积\")\n", "\n", "# 查看处理后的数据\n", "print(f\"\\n处理后数据形状: {df_processed.shape}\")\n", "df_processed.head()" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:38.031543Z", "iopub.status.busy": "2026-03-20T02:20:38.031291Z", "iopub.status.idle": "2026-03-20T02:20:38.751494Z", "shell.execute_reply": "2026-03-20T02:20:38.750808Z" } }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "

" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "与目标变量Y的相关性（绝对值排序）:\n", " Y: 1.0000\n", " p1: 0.3203\n", " p3_f: 0.3193\n", " p2: 0.0938\n", " gender_encoded: 0.0636\n", " age: 0.0521\n", " family_encoded: 0.0503\n" ] } ], "source": [ "# 特征相关性分析\n", "numeric_features = ['p1', 'p2', 'realty_ind', 'age', 'p3_f', \n", " 'gender_encoded', 'family_encoded', 'Y']\n", "correlation_matrix = df_processed[numeric_features].corr()\n", "\n", "plt.figure(figsize=(12, 10))\n", "sns.heatmap(correlation_matrix, \n", " annot=True, \n", " cmap='RdBu_r', \n", " center=0,\n", " square=True,\n", " fmt='.2f',\n", " cbar_kws={'label': '相关系数'})\n", "plt.title('特征相关性热力图', fontsize=16, fontweight='bold')\n", "plt.tight_layout()\n", "plt.savefig('correlation_heatmap.png', dpi=150, bbox_inches='tight')\n", "plt.show()\n", "\n", "# 打印与目标变量的相关性\n", "print(\"\\n与目标变量Y的相关性（绝对值排序）:\")\n", "target_corr = correlation_matrix['Y'].abs().sort_values(ascending=False)[:-1]\n", "for feature, corr in target_corr.items():\n", " print(f\" {feature}: {corr:.4f}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 4. 模型训练与评估" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:38.756346Z", "iopub.status.busy": "2026-03-20T02:20:38.756070Z", "iopub.status.idle": "2026-03-20T02:20:38.770384Z", "shell.execute_reply": "2026-03-20T02:20:38.769762Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "============================================================\n", "模型训练与评估\n", "============================================================\n", "\n", "数据划分:\n", " 训练集: 3240 样本 (80.0%)\n", " 测试集: 810 样本 (20.0%)\n", "\n", "训练集目标分布:\n", " Y=0: 2210 (68.2%)\n", " Y=1: 1030 (31.8%)\n" ] } ], "source": [ "# 准备特征和目标变量\n", "feature_cols = ['p1', 'p2', 'gender_encoded', 'family_encoded', \n", " 'realty_ind', 'age', 'p3_f']\n", "X = df_processed[feature_cols].values\n", "y = df_processed['Y'].values\n", "\n", "print(\"=\"*60)\n", "print(\"模型训练与评估\")\n", "print(\"=\"*60)\n", "\n", "# 划分训练集和测试集\n", "X_train, X_test, y_train, y_test = train_test_split(\n", " X, y, test_size=0.2, random_state=42, stratify=y\n", ")\n", "\n", "print(f\"\\n数据划分:\")\n", "print(f\" 训练集: {len(X_train)} 样本 ({len(X_train)/len(X)*100:.1f}%)\")\n", "print(f\" 测试集: {len(X_test)} 样本 ({len(X_test)/len(X)*100:.1f}%)\")\n", "print(f\"\\n训练集目标分布:\")\n", "print(f\" Y=0: {sum(y_train==0)} ({sum(y_train==0)/len(y_train)*100:.1f}%)\")\n", "print(f\" Y=1: {sum(y_train==1)} ({sum(y_train==1)/len(y_train)*100:.1f}%)\")" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:38.772955Z", "iopub.status.busy": "2026-03-20T02:20:38.772702Z", "iopub.status.idle": "2026-03-20T02:20:40.433782Z", "shell.execute_reply": "2026-03-20T02:20:40.432625Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "训练逻辑回归...\n", " ✓ 完成\n", " 准确率: 0.6617\n", " AUC: 0.6859\n", " 交叉验证: 0.6824 (+/- 0.0357)\n", "\n", "训练决策树...\n", " ✓ 完成\n", " 准确率: 0.6568\n", " AUC: 0.6931\n", " 交叉验证: 0.5926 (+/- 0.0408)\n", "\n", "训练随机森林...\n", " ✓ 完成\n", " 准确率: 0.6753\n", " AUC: 0.7002\n", " 交叉验证: 0.6802 (+/- 0.0412)\n" ] } ], "source": [ "# 定义模型\n", "models = {\n", " '逻辑回归': LogisticRegression(max_iter=1000, random_state=42, class_weight='balanced'),\n", " '决策树': DecisionTreeClassifier(max_depth=5, random_state=42, class_weight='balanced'),\n", " '随机森林': RandomForestClassifier(n_estimators=100, max_depth=10, random_state=42, class_weight='balanced')\n", "}\n", "\n", "# 训练并评估模型\n", "results = {}\n", "trained_models = {}\n", "\n", "for name, model in models.items():\n", " print(f\"\\n训练 {name}...\")\n", " \n", " # 训练模型\n", " model.fit(X_train, y_train)\n", " trained_models[name] = model\n", " \n", " # 预测\n", " y_pred = model.predict(X_test)\n", " y_prob = model.predict_proba(X_test)[:, 1]\n", " \n", " # 计算指标\n", " metrics = {\n", " 'accuracy': accuracy_score(y_test, y_pred),\n", " 'precision': precision_score(y_test, y_pred, zero_division=0),\n", " 'recall': recall_score(y_test, y_pred, zero_division=0),\n", " 'f1': f1_score(y_test, y_pred, zero_division=0),\n", " 'auc': roc_auc_score(y_test, y_prob),\n", " 'y_pred': y_pred,\n", " 'y_prob': y_prob\n", " }\n", " \n", " # 交叉验证\n", " cv_scores = cross_val_score(model, X_train, y_train, cv=5, scoring='accuracy')\n", " metrics['cv_mean'] = cv_scores.mean()\n", " metrics['cv_std'] = cv_scores.std()\n", " \n", " results[name] = metrics\n", " \n", " print(f\" ✓ 完成\")\n", " print(f\" 准确率: {metrics['accuracy']:.4f}\")\n", " print(f\" AUC: {metrics['auc']:.4f}\")\n", " print(f\" 交叉验证: {metrics['cv_mean']:.4f} (+/- {metrics['cv_std']*2:.4f})\")" ] }, { "cell_type": "code", "execution_count": 12, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:40.437759Z", "iopub.status.busy": "2026-03-20T02:20:40.437450Z", "iopub.status.idle": "2026-03-20T02:20:40.449097Z", "shell.execute_reply": "2026-03-20T02:20:40.448627Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "================================================================================\n", "模型性能对比\n", "================================================================================\n", " 模型准确率精确率召回率 F1分数 AUC CV均值\n", "逻辑回归 0.6617 0.4669 0.4380 0.4520 0.6859 0.6824\n", " 决策树 0.6568 0.4684 0.5736 0.5157 0.6931 0.5926\n", "随机森林 0.6753 0.4902 0.4845 0.4873 0.7002 0.6802\n", "\n", "最佳模型（基于F1分数）: 决策树\n", "F1分数: 0.5157\n" ] } ], "source": [ "# 模型对比表格\n", "print(\"\\n\" + \"=\"*80)\n", "print(\"模型性能对比\")\n", "print(\"=\"*80)\n", "\n", "comparison_df = pd.DataFrame({\n", " '模型': list(results.keys()),\n", " '准确率': [f\"{r['accuracy']:.4f}\" for r in results.values()],\n", " '精确率': [f\"{r['precision']:.4f}\" for r in results.values()],\n", " '召回率': [f\"{r['recall']:.4f}\" for r in results.values()],\n", " 'F1分数': [f\"{r['f1']:.4f}\" for r in results.values()],\n", " 'AUC': [f\"{r['auc']:.4f}\" for r in results.values()],\n", " 'CV均值': [f\"{r['cv_mean']:.4f}\" for r in results.values()]\n", "})\n", "\n", "print(comparison_df.to_string(index=False))\n", "\n", "# 找出最佳模型\n", "best_model_name = max(results.items(), key=lambda x: x[1]['f1'])[0]\n", "print(f\"\\n最佳模型（基于F1分数）: {best_model_name}\")\n", "print(f\"F1分数: {results[best_model_name]['f1']:.4f}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 5. 结果可视化" ] }, { "cell_type": "code", "execution_count": 13, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:40.451816Z", "iopub.status.busy": "2026-03-20T02:20:40.451551Z", "iopub.status.idle": "2026-03-20T02:20:43.319965Z", "shell.execute_reply": "2026-03-20T02:20:43.318354Z" } }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "

" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "✅ 综合分析图表已生成并保存\n" ] } ], "source": [ "# 创建综合可视化图\n", "fig = plt.figure(figsize=(20, 16))\n", "\n", "# 创建GridSpec布局\n", "gs = fig.add_gridspec(3, 3, hspace=0.3, wspace=0.3)\n", "\n", "# 1. 模型性能对比柱状图\n", "ax1 = fig.add_subplot(gs[0, 0])\n", "model_names = list(results.keys())\n", "accuracies = [results[m]['accuracy'] for m in model_names]\n", "colors_models = ['#FF6B6B', '#4ECDC4', '#45B7D1']\n", "bars = ax1.bar(model_names, accuracies, color=colors_models, alpha=0.8, edgecolor='black')\n", "ax1.set_ylabel('准确率', fontsize=12)\n", "ax1.set_title('模型准确率对比', fontsize=14, fontweight='bold')\n", "ax1.set_ylim(0, 1)\n", "ax1.grid(axis='y', alpha=0.3)\n", "for bar, acc in zip(bars, accuracies):\n", " ax1.text(bar.get_x() + bar.get_width()/2, bar.get_height() + 0.02, \n", " f'{acc:.3f}', ha='center', fontsize=11, fontweight='bold')\n", "\n", "# 2. 多指标雷达图/柱状图\n", "ax2 = fig.add_subplot(gs[0, 1])\n", "metrics_names = ['准确率', '精确率', '召回率', 'F1', 'AUC']\n", "x = np.arange(len(metrics_names))\n", "width = 0.25\n", "for i, (name, res) in enumerate(results.items()):\n", " values = [res['accuracy'], res['precision'], res['recall'], res['f1'], res['auc']]\n", " ax2.bar(x + i*width, values, width, label=name, color=colors_models[i], alpha=0.8)\n", "ax2.set_xticks(x + width)\n", "ax2.set_xticklabels(metrics_names, fontsize=10)\n", "ax2.set_ylabel('分数', fontsize=12)\n", "ax2.set_title('多指标性能对比', fontsize=14, fontweight='bold')\n", "ax2.legend(fontsize=10)\n", "ax2.set_ylim(0, 1)\n", "ax2.grid(axis='y', alpha=0.3)\n", "\n", "# 3. ROC曲线\n", "ax3 = fig.add_subplot(gs[0, 2])\n", "for i, (name, res) in enumerate(results.items()):\n", " fpr, tpr, _ = roc_curve(y_test, res['y_prob'])\n", " ax3.plot(fpr, tpr, color=colors_models[i], linewidth=2, \n", " label=f\"{name} (AUC={res['auc']:.3f})\")\n", "ax3.plot([0, 1], [0, 1], 'k--', linewidth=1, alpha=0.5)\n", "ax3.set_xlabel('假阳性率 (False Positive Rate)', fontsize=11)\n", "ax3.set_ylabel('真阳性率 (True Positive Rate)', fontsize=11)\n", "ax3.set_title('ROC曲线对比', fontsize=14, fontweight='bold')\n", "ax3.legend(fontsize=9, loc='lower right')\n", "ax3.grid(alpha=0.3)\n", "\n", "# 4. 混淆矩阵 - 逻辑回归\n", "ax4 = fig.add_subplot(gs[1, 0])\n", "cm_lr = confusion_matrix(y_test, results['逻辑回归']['y_pred'])\n", "sns.heatmap(cm_lr, annot=True, fmt='d', cmap='Blues', ax=ax4, cbar=False)\n", "ax4.set_title('混淆矩阵 - 逻辑回归', fontsize=14, fontweight='bold')\n", "ax4.set_xlabel('预测标签', fontsize=11)\n", "ax4.set_ylabel('真实标签', fontsize=11)\n", "ax4.set_xticklabels(['0', '1'])\n", "ax4.set_yticklabels(['0', '1'])\n", "\n", "# 5. 混淆矩阵 - 决策树\n", "ax5 = fig.add_subplot(gs[1, 1])\n", "cm_dt = confusion_matrix(y_test, results['决策树']['y_pred'])\n", "sns.heatmap(cm_dt, annot=True, fmt='d', cmap='Greens', ax=ax5, cbar=False)\n", "ax5.set_title('混淆矩阵 - 决策树', fontsize=14, fontweight='bold')\n", "ax5.set_xlabel('预测标签', fontsize=11)\n", "ax5.set_ylabel('真实标签', fontsize=11)\n", "ax5.set_xticklabels(['0', '1'])\n", "ax5.set_yticklabels(['0', '1'])\n", "\n", "# 6. 混淆矩阵 - 随机森林\n", "ax6 = fig.add_subplot(gs[1, 2])\n", "cm_rf = confusion_matrix(y_test, results['随机森林']['y_pred'])\n", "sns.heatmap(cm_rf, annot=True, fmt='d', cmap='Oranges', ax=ax6, cbar=False)\n", "ax6.set_title('混淆矩阵 - 随机森林', fontsize=14, fontweight='bold')\n", "ax6.set_xlabel('预测标签', fontsize=11)\n", "ax6.set_ylabel('真实标签', fontsize=11)\n", "ax6.set_xticklabels(['0', '1'])\n", "ax6.set_yticklabels(['0', '1'])\n", "\n", "# 7. 特征重要性（随机森林）\n", "ax7 = fig.add_subplot(gs[2, 0])\n", "rf_model = trained_models['随机森林']\n", "importance_df = pd.DataFrame({\n", " 'feature': feature_cols,\n", " 'importance': rf_model.feature_importances_\n", "}).sort_values('importance', ascending=True)\n", "colors_feat = plt.cm.viridis(np.linspace(0, 1, len(feature_cols)))\n", "ax7.barh(importance_df['feature'], importance_df['importance'], color=colors_feat)\n", "ax7.set_xlabel('重要性', fontsize=11)\n", "ax7.set_title('特征重要性（随机森林）', fontsize=14, fontweight='bold')\n", "ax7.grid(axis='x', alpha=0.3)\n", "\n", "# 8. 预测概率分布\n", "ax8 = fig.add_subplot(gs[2, 1])\n", "best_pred = results[best_model_name]['y_prob']\n", "ax8.hist(best_pred[y_test==0], bins=30, alpha=0.6, label='真实标签=0', color='coral')\n", "ax8.hist(best_pred[y_test==1], bins=30, alpha=0.6, label='真实标签=1', color='skyblue')\n", "ax8.axvline(x=0.5, color='red', linestyle='--', linewidth=2, label='阈值=0.5')\n", "ax8.set_xlabel('预测概率', fontsize=11)\n", "ax8.set_ylabel('频数', fontsize=11)\n", "ax8.set_title(f'预测概率分布 - {best_model_name}', fontsize=14, fontweight='bold')\n", "ax8.legend(fontsize=9)\n", "ax8.grid(alpha=0.3)\n", "\n", "# 9. 交叉验证分数\n", "ax9 = fig.add_subplot(gs[2, 2])\n", "cv_means = [results[m]['cv_mean'] for m in model_names]\n", "cv_stds = [results[m]['cv_std'] for m in model_names]\n", "bars = ax9.bar(model_names, cv_means, yerr=cv_stds, color=colors_models, \n", " alpha=0.8, capsize=5, edgecolor='black')\n", "ax9.set_ylabel('交叉验证准确率', fontsize=12)\n", "ax9.set_title('5折交叉验证结果', fontsize=14, fontweight='bold')\n", "ax9.set_ylim(0, 1)\n", "ax9.grid(axis='y', alpha=0.3)\n", "for bar, mean, std in zip(bars, cv_means, cv_stds):\n", " ax9.text(bar.get_x() + bar.get_width()/2, bar.get_height() + std + 0.02, \n", " f'{mean:.3f}', ha='center', fontsize=10, fontweight='bold')\n", "\n", "plt.suptitle('机器学习分类模型分析结果总览', fontsize=18, fontweight='bold', y=0.995)\n", "plt.savefig('ml_analysis_results.png', dpi=200, bbox_inches='tight')\n", "plt.show()\n", "\n", "print(\"✅ 综合分析图表已生成并保存\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 6. 详细分类报告" ] }, { "cell_type": "code", "execution_count": 14, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:43.330873Z", "iopub.status.busy": "2026-03-20T02:20:43.330593Z", "iopub.status.idle": "2026-03-20T02:20:43.343180Z", "shell.execute_reply": "2026-03-20T02:20:43.342629Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "============================================================\n", "详细分类报告 - 决策树\n", "============================================================\n", " precision recall f1-score support\n", "\n", " Class 0 0.78 0.70 0.73 552\n", " Class 1 0.47 0.57 0.52 258\n", "\n", " accuracy 0.66 810\n", " macro avg 0.62 0.63 0.62 810\n", "weighted avg 0.68 0.66 0.66 810\n", "\n", "\n", "其他指标:\n", " 特异度 (Specificity): 0.6957\n", " 敏感度 (Sensitivity): 0.5736\n", " 真阴性: 384, 假阳性: 168, 假阴性: 110, 真阳性: 148\n" ] } ], "source": [ "# 输出最佳模型的详细分类报告\n", "print(\"=\"*60)\n", "print(f\"详细分类报告 - {best_model_name}\")\n", "print(\"=\"*60)\n", "\n", "best_predictions = results[best_model_name]['y_pred']\n", "print(classification_report(y_test, best_predictions, \n", " target_names=['Class 0', 'Class 1']))\n", "\n", "# 计算其他指标\n", "tn, fp, fn, tp = confusion_matrix(y_test, best_predictions).ravel()\n", "specificity = tn / (tn + fp) # 特异度\n", "sensitivity = tp / (tp + fn) # 敏感度（召回率）\n", "\n", "print(f\"\\n其他指标:\")\n", "print(f\" 特异度 (Specificity): {specificity:.4f}\")\n", "print(f\" 敏感度 (Sensitivity): {sensitivity:.4f}\")\n", "print(f\" 真阴性: {tn}, 假阳性: {fp}, 假阴性: {fn}, 真阳性: {tp}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 7. 模型可解释性分析" ] }, { "cell_type": "code", "execution_count": 15, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:43.345609Z", "iopub.status.busy": "2026-03-20T02:20:43.345405Z", "iopub.status.idle": "2026-03-20T02:20:43.370791Z", "shell.execute_reply": "2026-03-20T02:20:43.370197Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "============================================================\n", "特征重要性分析 - 随机森林\n", "============================================================\n", "\n", "特征重要性排序:\n", " 特征名重要性重要性%\n", " age 0.338661 33.866113\n", " p3_f 0.311285 31.128494\n", " p1 0.163532 16.353204\n", "family_encoded 0.097871 9.787128\n", " realty_ind 0.034668 3.466797\n", "gender_encoded 0.033615 3.361490\n", " p2 0.020368 2.036774\n", "\n", "============================================================\n", "逻辑回归系数分析\n", "============================================================\n", " 特征名系数影响\n", " p1 1.177351 正向\n", " p2 0.957652 正向\n", "gender_encoded 0.255301 正向\n", " realty_ind 0.092006 正向\n", " p3_f 0.065485 正向\n", " age -0.009538 负向\n", "family_encoded -0.150484 负向\n", "\n", "截距 (Intercept): -1.1569\n" ] } ], "source": [ "# 特征重要性详细分析\n", "print(\"=\"*60)\n", "print(\"特征重要性分析 - 随机森林\")\n", "print(\"=\"*60)\n", "\n", "feature_importance = pd.DataFrame({\n", " '特征名': feature_cols,\n", " '重要性': rf_model.feature_importances_,\n", " '重要性%': rf_model.feature_importances_ * 100\n", "}).sort_values('重要性', ascending=False)\n", "\n", "print(\"\\n特征重要性排序:\")\n", "print(feature_importance.to_string(index=False))\n", "\n", "# 逻辑回归系数\n", "print(\"\\n\" + \"=\"*60)\n", "print(\"逻辑回归系数分析\")\n", "print(\"=\"*60)\n", "\n", "lr_model = trained_models['逻辑回归']\n", "lr_coef = pd.DataFrame({\n", " '特征名': feature_cols,\n", " '系数': lr_model.coef_[0],\n", " '影响': ['正向' if c > 0 else '负向' for c in lr_model.coef_[0]]\n", "}).sort_values('系数', ascending=False)\n", "\n", "print(lr_coef.to_string(index=False))\n", "print(f\"\\n截距 (Intercept): {lr_model.intercept_[0]:.4f}\")" ] }, { "cell_type": "code", "execution_count": 16, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:43.373099Z", "iopub.status.busy": "2026-03-20T02:20:43.372873Z", "iopub.status.idle": "2026-03-20T02:20:44.163363Z", "shell.execute_reply": "2026-03-20T02:20:44.162354Z" } }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "

" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# 可视化特征重要性对比\n", "fig, axes = plt.subplots(1, 2, figsize=(16, 6))\n", "\n", "# 随机森林特征重要性\n", "ax1 = axes[0]\n", "importance_sorted = feature_importance.sort_values('重要性', ascending=True)\n", "colors = plt.cm.RdYlGn(np.linspace(0.2, 0.8, len(feature_cols)))\n", "bars = ax1.barh(importance_sorted['特征名'], importance_sorted['重要性'], color=colors)\n", "ax1.set_xlabel('重要性', fontsize=12)\n", "ax1.set_title('随机森林 - 特征重要性', fontsize=14, fontweight='bold')\n", "ax1.grid(axis='x', alpha=0.3)\n", "for i, (bar, val) in enumerate(zip(bars, importance_sorted['重要性'])):\n", " ax1.text(bar.get_width() + 0.005, bar.get_y() + bar.get_height()/2, \n", " f'{val:.3f}', va='center', fontsize=10)\n", "\n", "# 逻辑回归系数\n", "ax2 = axes[1]\n", "lr_coef_sorted = lr_coef.sort_values('系数', ascending=True)\n", "colors_lr = ['red' if c < 0 else 'green' for c in lr_coef_sorted['系数']]\n", "bars2 = ax2.barh(lr_coef_sorted['特征名'], lr_coef_sorted['系数'], color=colors_lr, alpha=0.7)\n", "ax2.set_xlabel('系数值', fontsize=12)\n", "ax2.set_title('逻辑回归 - 特征系数', fontsize=14, fontweight='bold')\n", "ax2.axvline(x=0, color='black', linestyle='-', linewidth=0.8)\n", "ax2.grid(axis='x', alpha=0.3)\n", "for i, (bar, val) in enumerate(zip(bars2, lr_coef_sorted['系数'])):\n", " offset = 0.05 if val > 0 else -0.15\n", " ax2.text(val + offset, bar.get_y() + bar.get_height()/2, \n", " f'{val:.3f}', va='center', fontsize=10)\n", "\n", "plt.tight_layout()\n", "plt.savefig('feature_importance_analysis.png', dpi=150, bbox_inches='tight')\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 8. 业务洞察与结论" ] }, { "cell_type": "code", "execution_count": 17, "metadata": { "execution": { "iopub.execute_input": "2026-03-20T02:20:44.171579Z", "iopub.status.busy": "2026-03-20T02:20:44.171196Z", "iopub.status.idle": "2026-03-20T02:20:44.190095Z", "shell.execute_reply": "2026-03-20T02:20:44.189626Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "======================================================================\n", "机器学习分类分析 - 业务洞察报告\n", "======================================================================\n", "\n", "📊 数据概况:\n", " • 总样本数: 4,050\n", " • 特征数量: 7\n", " • 目标变量分布: Y=0 (68.2%), Y=1 (31.8%)\n", "\n", "🏆 最佳模型: 决策树\n", " • 准确率: 65.68%\n", " • F1分数: 0.5157\n", " • AUC: 0.6931\n", "\n", "🔍 关键发现:\n", " • 最重要的3个特征: age, p3_f, p1\n", " • 目标变量比例最高的年龄组: 30-40 (34.74%)\n", " • M性目标变量比例更高 (35.90% vs 29.67%)\n", " • 有房产者目标变量比例: 32.04%\n", " • 无房产者目标变量比例: 31.27%\n", "\n", "💡 建议:\n", " 1. 重点关注特征: age、p3_f、p1\n", " 2. 针对M性群体制定针对性策略\n", " 3. 年龄组30-40是重点目标人群\n", " 4. 模型可用于预测，建议定期重训练以保持准确性\n", "\n", "======================================================================\n", "分析完成！所有图表已保存。\n", "======================================================================\n" ] } ], "source": [ "# 生成分析报告\n", "print(\"=\"*70)\n", "print(\"机器学习分类分析 - 业务洞察报告\")\n", "print(\"=\"*70)\n", "\n", "print(f\"\\n📊 数据概况:\")\n", "print(f\" • 总样本数: {len(df):,}\")\n", "print(f\" • 特征数量: {len(feature_cols)}\")\n", "print(f\" • 目标变量分布: Y=0 ({sum(df['Y']==0)/len(df)*100:.1f}%), Y=1 ({sum(df['Y']==1)/len(df)*100:.1f}%)\")\n", "\n", "print(f\"\\n🏆 最佳模型: {best_model_name}\")\n", "print(f\" • 准确率: {results[best_model_name]['accuracy']:.2%}\")\n", "print(f\" • F1分数: {results[best_model_name]['f1']:.4f}\")\n", "print(f\" • AUC: {results[best_model_name]['auc']:.4f}\")\n", "\n", "print(f\"\\n🔍 关键发现:\")\n", "\n", "# 最重要的3个特征\n", "top3_features = feature_importance.head(3)['特征名'].tolist()\n", "print(f\" • 最重要的3个特征: {', '.join(top3_features)}\")\n", "\n", "# 年龄分析\n", "age_group_analysis = df_processed.groupby('age_group')['Y'].mean().sort_values(ascending=False)\n", "print(f\" • 目标变量比例最高的年龄组: {age_group_analysis.index[0]} ({age_group_analysis.iloc[0]:.2%})\")\n", "\n", "# 性别分析\n", "gender_analysis = df_processed.groupby('gender')['Y'].mean()\n", "higher_gender = gender_analysis.idxmax()\n", "print(f\" • {higher_gender}性目标变量比例更高 ({gender_analysis.max():.2%} vs {gender_analysis.min():.2%})\")\n", "\n", "# 房产分析\n", "realty_analysis = df_processed.groupby('realty_ind')['Y'].mean()\n", "print(f\" • 有房产者目标变量比例: {realty_analysis[1]:.2%}\")\n", "print(f\" • 无房产者目标变量比例: {realty_analysis[0]:.2%}\")\n", "\n", "print(f\"\\n💡 建议:\")\n", "print(f\" 1. 重点关注特征: {top3_features[0]}、{top3_features[1]}、{top3_features[2]}\")\n", "print(f\" 2. 针对{higher_gender}性群体制定针对性策略\")\n", "print(f\" 3. 年龄组{age_group_analysis.index[0]}是重点目标人群\")\n", "print(f\" 4. 模型可用于预测，建议定期重训练以保持准确性\")\n", "\n", "print(\"\\n\" + \"=\"*70)\n", "print(\"分析完成！所有图表已保存。\")\n", "print(\"=\"*70)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "---\n", "\n", "## 📁 生成的文件\n", "\n", "本Notebook运行后会生成以下文件：\n", "- `eda_overview.png` - 探索性数据分析综合图\n", "- `correlation_heatmap.png` - 特征相关性热力图\n", "- `ml_analysis_results.png` - 机器学习分析结果总览（9宫格）\n", "- `feature_importance_analysis.png` - 特征重要性对比分析\n", "\n", "## 📝 总结\n", "\n", "本Notebook演示了完整的机器学习数据分析流程：\n", "1. **数据加载与清洗** - 数据质量检查和预处理\n", "2. **探索性数据分析** - 统计摘要和可视化\n", "3. **特征工程** - 编码、分箱和交互特征\n", "4. **模型训练** - 3种分类算法对比\n", "5. **模型评估** - 多维度指标和交叉验证\n", "6. **结果可视化** - 9宫格综合展示\n", "7. **可解释性分析** - 特征重要性和系数解读\n", "8. **业务洞察** - 结合业务场景的建议" ] } ], "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": "3.9.12" } }, "nbformat": 4, "nbformat_minor": 4 }