AdaBoost算法之葡萄酒案例

AdaBoostClassifier(
    n_estimators=500,    # 500棵树
    learning_rate=0.1,    # 学习率
    random_state=22       # 随机种子
)

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import AdaBoostClassifier
from sklearn.metrics import accuracy_score
from sklearn.preprocessing import LabelEncoder
from sklearn.datasets import load_wine

# 1. 加载葡萄酒数据
def load_data():
    """加载葡萄酒数据"""
    wine = load_wine()
    data = pd.DataFrame(wine.data, columns=wine.feature_names)
    data['target'] = wine.target
    
    print(f"数据形状: {data.shape}")
    print(f"标签分布:\n{data['target'].value_counts()}")
    print(f"\n前5行:")
    print(data.head())
    
    return data

# 2. 数据预处理
def preprocess_data(data):
    """数据预处理"""
    # 2.1 去掉一个类别（删除类别0，保留1和2）
    data = data[data['target'] != 0].copy()
    print(f"\n删除类别0后: {data.shape}")
    print(f"标签分布:\n{data['target'].value_counts()}")
    
    # 2.2 提取特征（酒精含量和色泽度）
    X = data[['alcohol', 'hue']]
    y = data['target']
    
    # 2.3 标签编码（将1,2转为0,1）
    le = LabelEncoder()
    y = le.fit_transform(y)
    
    print(f"\n编码后标签: {np.unique(y)}")
    print(f"特征形状: {X.shape}")
    
    return X, y, le

# 3. 划分数据集
def split_data(X, y):
    """划分训练集和测试集"""
    X_train, X_test, y_train, y_test = train_test_split(
        X, y, test_size=0.2, random_state=22
    )
    print(f"\n训练集: {X_train.shape}")
    print(f"测试集: {X_test.shape}")
    return X_train, X_test, y_train, y_test

# 4. 单一决策树
def train_decision_tree(X_train, X_test, y_train, y_test):
    """单一决策树"""
    print("\n=== 单一决策树 ===")
    dt = DecisionTreeClassifier(
        criterion='entropy',  # 使用熵
        max_depth=5,
        random_state=22
    )
    dt.fit(X_train, y_train)
    
    y_pred = dt.predict(X_test)
    acc = accuracy_score(y_test, y_pred)
    print(f"准确率: {acc:.4f}")
    return dt, acc

# 5. AdaBoost
def train_adaboost(X_train, X_test, y_train, y_test):
    """AdaBoost"""
    print("\n=== AdaBoost ===")
    # 底层决策树
    dt = DecisionTreeClassifier(
        criterion='entropy',
        max_depth=5,
        random_state=22
    )
    
    # AdaBoost
    ab = AdaBoostClassifier(
        estimator=dt,  # 底层决策树
        n_estimators=500,
        learning_rate=0.1,
        random_state=22
    )
    ab.fit(X_train, y_train)
    
    y_pred = ab.predict(X_test)
    acc = accuracy_score(y_test, y_pred)
    print(f"准确率: {acc:.4f}")
    return ab, acc

# 6. 主函数
def main():
    """主函数"""
    print("=" * 50)
    print("AdaBoost - 葡萄酒分类案例")
    print("=" * 50)
    
    # 1. 加载数据
    data = load_data()
    
    # 2. 数据预处理
    X, y, le = preprocess_data(data)
    
    # 3. 划分数据集
    X_train, X_test, y_train, y_test = split_data(X, y)
    
    # 4. 单一决策树
    dt, dt_acc = train_decision_tree(X_train, X_test, y_train, y_test)
    
    # 5. AdaBoost
    ab, ab_acc = train_adaboost(X_train, X_test, y_train, y_test)
    
    # 6. 对比
    print("\n=== 对比 ===")
    print(f"单一决策树: {dt_acc:.4f}")
    print(f"AdaBoost: {ab_acc:.4f}")
    
    # 7. 预测新样本
    print("\n=== 预测新样本 ===")
    samples = [
        {'alcohol': 13.5, 'hue': 1.05},
        {'alcohol': 12.0, 'hue': 0.80},
        {'alcohol': 14.0, 'hue': 1.10}
    ]
    
    X_new = pd.DataFrame(samples)
    predictions = ab.predict(X_new)
    
    for sample, pred in zip(samples, predictions):
        label = le.inverse_transform([pred])[0]
        print(f"{sample}: 类别{label}")

if __name__ == '__main__':
    main()