GBDT / XGBoost 的本质不是“树多“：安全攻防视角下的梯度提升原理

from sklearn.ensemble import GradientBoostingClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix
import numpy as np

# 生成模拟安全数据（恶意软件分类）
np.random.seed(42)
X = np.random.rand(1000, 20)  # 20个特征
y = np.random.randint(0, 2, 1000)  # 二分类标签

# 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=42)

# 初始化GBDT分类器
gbdt = GradientBoostingClassifier(
    n_estimators=100,  # 树的数量
    learning_rate=0.1,  # 学习率
    max_depth=3,  # 树的最大深度
    random_state=42
)

# 训练模型
gbdt.fit(X_train, y_train)

# 预测
y_pred = gbdt.predict(X_test)

# 评估模型
print("分类报告：")
print(classification_report(y_test, y_pred))
print("混淆矩阵：")
print(confusion_matrix(y_test, y_pred))

分类报告：
              precision    recall  f1-score   support

           0       0.52      0.51      0.52       102
           1       0.49      0.50      0.49        98

    accuracy                           0.50       200
   macro avg       0.50      0.50      0.50       200
weighted avg       0.50      0.50      0.50       200

混淆矩阵：
[[52 50]
 [49 49]]

import xgboost as xgb
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix
import numpy as np

# 生成模拟安全数据（入侵检测）
np.random.seed(42)
X = np.random.rand(1000, 20)  # 20个特征
y = np.random.randint(0, 2, 1000)  # 二分类标签

# 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=42)

# 转换为DMatrix格式（XGBoost专用格式）
train_dmatrix = xgb.DMatrix(X_train, label=y_train)
test_dmatrix = xgb.DMatrix(X_test, label=y_test)

# 设置XGBoost参数
params = {
    'objective': 'binary:logistic',  # 二分类任务
    'max_depth': 3,  # 树的最大深度
    'learning_rate': 0.1,  # 学习率
    'n_estimators': 100,  # 树的数量
    'alpha': 0.1,  # L1正则化参数
    'lambda': 0.1,  # L2正则化参数
    'random_state': 42
}

# 训练模型
xgb_model = xgb.train(params, train_dmatrix, num_boost_round=100)

# 预测
y_pred = xgb_model.predict(test_dmatrix)
# 将概率转换为类别
y_pred_class = [1 if prob > 0.5 else 0 for prob in y_pred]

# 评估模型
print("分类报告：")
print(classification_report(y_test, y_pred_class))
print("混淆矩阵：")
print(confusion_matrix(y_test, y_pred_class))

分类报告：
              precision    recall  f1-score   support

           0       0.53      0.52      0.53       102
           1       0.50      0.51      0.50        98

    accuracy                           0.52       200
   macro avg       0.51      0.51      0.51       200
weighted avg       0.52      0.52      0.52       200

混淆矩阵：
[[53 49]
 [48 50]]

import xgboost as xgb
from sklearn.model_selection import GridSearchCV, train_test_split
from sklearn.metrics import f1_score
import numpy as np

# 生成模拟安全数据（DDoS攻击预测）
np.random.seed(42)
X = np.random.rand(1000, 20)  # 20个特征
y = np.random.randint(0, 2, 1000)  # 二分类标签

# 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=42)

# 初始化XGBoost分类器
xgb_clf = xgb.XGBClassifier(
    objective='binary:logistic',
    random_state=42
)

# 定义超参数搜索空间
param_grid = {
    'n_estimators': [50, 100, 150],
    'learning_rate': [0.01, 0.1, 0.3],
    'max_depth': [3, 5, 7],
    'alpha': [0, 0.1, 0.5],
    'lambda': [0, 0.1, 0.5]
}

# 网格搜索调优
grid_search = GridSearchCV(
    estimator=xgb_clf,
    param_grid=param_grid,
    scoring='f1',  # 以F1分数作为评估指标
    cv=5,  # 5折交叉验证
    n_jobs=-1  # 使用所有CPU核心
)

# 训练和调优
grid_search.fit(X_train, y_train)

# 输出最佳参数
print("最佳参数：")
print(grid_search.best_params_)

# 使用最佳模型预测
best_model = grid_search.best_estimator_
y_pred = best_model.predict(X_test)

# 评估模型
print("最佳模型F1分数：")
print(f1_score(y_test, y_pred))

最佳参数：
{'alpha': 0.1, 'lambda': 0.5, 'learning_rate': 0.3, 'max_depth': 5, 'n_estimators': 150}
最佳模型F1分数：
0.5411764705882353

import xgboost as xgb
import matplotlib.pyplot as plt
from sklearn.datasets import make_classification
from sklearn.model_selection import train_test_split

# 生成模拟安全数据
X, y = make_classification(n_samples=1000, n_features=20, n_informative=10, random_state=42)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# 训练XGBoost模型
model = xgb.XGBClassifier(n_estimators=100, max_depth=3, learning_rate=0.1, random_state=42)
model.fit(X_train, y_train)

# 获取特征重要性
feature_importance = model.feature_importances_

# 可视化特征重要性
plt.figure(figsize=(10, 6))
plt.bar(range(len(feature_importance)), feature_importance)
plt.title('Feature Importance - XGBoost')
plt.xlabel('Feature Index')
plt.ylabel('Importance Score')
plt.xticks(range(len(feature_importance)))
plt.grid(True, alpha=0.3)
plt.show()

# 安装所需库
pip install numpy pandas scikit-learn xgboost matplotlib seaborn

# 验证安装
python -c "import xgboost; print(xgboost.__version__)"