如何在GridSearchCV中应用Dropout

from keras.models import Sequential
from keras.layers import Dense, Dropout
from keras.wrappers.scikit_learn import KerasClassifier
from sklearn.model_selection import GridSearchCV
from sklearn.datasets import make_classification
from keras import backend as K

# 定义自定义的Estimator类
class DropoutEstimator(KerasClassifier):
    def __init__(self, **kwargs):
        super().__init__(**kwargs)
    
    def fit(self, X, y):
        K.set_learning_phase(1)  # 启用Dropout
        super().fit(X, y)
        K.set_learning_phase(0)  # 禁用Dropout
    
    def predict(self, X):
        K.set_learning_phase(0)  # 禁用Dropout
        return super().predict(X)

# 创建一个简单的神经网络模型
def create_model(dropout_rate=0.0):
    model = Sequential()
    model.add(Dense(64, activation='relu', input_dim=20))
    model.add(Dropout(dropout_rate))
    model.add(Dense(1, activation='sigmoid'))
    model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
    return model

# 创建数据集
X, y = make_classification(n_samples=1000, n_features=20, random_state=42)

# 创建DropoutEstimator实例
estimator = DropoutEstimator(build_fn=create_model, verbose=0)

# 定义参数网格
param_grid = {
    'dropout_rate': [0.0, 0.2, 0.4, 0.6, 0.8]
}

# 创建GridSearchCV实例
grid_search = GridSearchCV(estimator=estimator, param_grid=param_grid, cv=5)

# 执行网格搜索和交叉验证
grid_search.fit(X, y)

# 输出最佳参数和得分
print("Best parameters: ", grid_search.best_params_)
print("Best score: ", grid_search.best_score_)