RF分类器（二分类）超参数贝叶斯优化 python代码

以下是使用贝叶斯优化调参的RF分类器的Python代码示例：

from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import roc_auc_score
from bayes_opt import BayesianOptimization

# 定义优化函数
def rf_cv(n_estimators, max_depth, min_samples_split, 
          min_samples_leaf, max_features, data, targets):
    estimator = RandomForestClassifier(
        n_estimators=n_estimators,
        max_depth=max_depth,
        min_samples_split=min_samples_split,
        min_samples_leaf=min_samples_leaf,
        max_features=max_features,
        random_state=2,
        n_jobs=-1
    )
    cval = cross_val_score(estimator, data, targets, 
                           scoring='roc_auc', cv=5)
    return cval.mean()

# 定义超参数的搜索范围
rf_bo = BayesianOptimization(
    rf_cv,
    {
        'n_estimators': (10, 250),
        'max_depth': (5, 15),
        'min_samples_split': (2, 25),
        'min_samples_leaf': (1, 25),
        'max_features': (1, 10)
    },
    random_state=2
)

# 进行贝叶斯优化，寻求最优超参数
rf_bo.maximize(n_iter=10, init_points=5)

# 使用最优超参数构建RF分类器并进行预测
best_params = rf_bo.max['params']
rf_clf = RandomForestClassifier(
    n_estimators=int(best_params['n_estimators']),
    max_depth=int(best_params['max_depth']),
    min_samples_split=int(best_params['min_samples_split']),
    min_samples_leaf=int(best_params['min_samples_leaf']),
    max_features=int(best_params['max_features']),
    random_state=2,
    n_jobs=-1
)

rf_clf.fit(X_train, y_train)
y_pred = rf_clf.predict(X_test)
auc_score = roc_auc_score(y_test, y_pred)
print('AUC score:', auc_score)

请注意，这个示例中的超参数搜索范围是手动选择的，并且不一定适用于您的数据集。您可能需要根据您的数据集调整搜索范围。