集成学习之大纲介绍

import numpy as np
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.datasets import make_classification

# 1. 生成数据
X, y = make_classification(
    n_samples=1000,
    n_features=20,
    n_informative=10,
    n_classes=2,
    random_state=42
)

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

print(f"训练集: {X_train.shape}")
print(f"测试集: {X_test.shape}")

# 2. 单个决策树（弱学习器）
print("\n=== 单个决策树（弱学习器）===")
dt = DecisionTreeClassifier(random_state=42)
dt.fit(X_train, y_train)
dt_acc = accuracy_score(y_test, dt.predict(X_test))
print(f"准确率: {dt_acc:.4f}")

# 3. 随机森林（Bagging思想）
print("\n=== 随机森林（Bagging）===")
rf = RandomForestClassifier(
    n_estimators=100,  # 100棵决策树
    random_state=42
)
rf.fit(X_train, y_train)
rf_acc = accuracy_score(y_test, rf.predict(X_test))
print(f"准确率: {rf_acc:.4f}")

# 4. AdaBoost（Boosting思想）
print("\n=== AdaBoost（Boosting）===")
ab = AdaBoostClassifier(
    n_estimators=100,
    random_state=42
)
ab.fit(X_train, y_train)
ab_acc = accuracy_score(y_test, ab.predict(X_test))
print(f"准确率: {ab_acc:.4f}")

# 5. 对比
print("\n=== 对比 ===")
print(f"单个决策树: {dt_acc:.4f}")
print(f"随机森林(Bagging): {rf_acc:.4f}")
print(f"AdaBoost(Boosting): {ab_acc:.4f}")

# 6. 模拟Bagging的有放回抽样
print("\n=== 模拟Bagging有放回抽样 ===")
np.random.seed(42)
data_size = 1000
sample_size = 500

# 三个模型各自抽样
sample1 = np.random.choice(data_size, sample_size, replace=True)
sample2 = np.random.choice(data_size, sample_size, replace=True)
sample3 = np.random.choice(data_size, sample_size, replace=True)

# 计算交集
intersection_12 = len(set(sample1) & set(sample2))
intersection_13 = len(set(sample1) & set(sample3))
intersection_23 = len(set(sample2) & set(sample3))

print(f"模型1和模型2的交集: {intersection_12}")
print(f"模型1和模型3的交集: {intersection_13}")
print(f"模型2和模型3的交集: {intersection_23}")
print("既有交集又有差异，投票才有意义")

# 7. 投票决策示例
print("\n=== 投票决策示例 ===")
# 三个模型的预测结果
pred1 = np.array([1, 0, 1, 1, 0])
pred2 = np.array([1, 1, 1, 0, 0])
pred3 = np.array([0, 1, 1, 1, 0])

# 投票（多数表决）
from scipy.stats import mode
predictions = np.array([pred1, pred2, pred3])
final_pred = mode(predictions, axis=0)[0].ravel()

print(f"模型1: {pred1}")
print(f"模型2: {pred2}")
print(f"模型3: {pred3}")
print(f"投票结果: {final_pred}")