从 keras.wrappers.sci-learn 迁移到 scikeras

Question

我在编程方面不是很有经验，所以我真的不知道如何进行迁移，因为没有太多关于它的文档，我唯一找到的是这个网站https://adriangb.com/scikeras/stable/migration.html 我也尝试回到旧的 keras.wrappers 但它不再可用

任何指导或帮助将不胜感激。

这是我的代码

import tensorflow as tf
from sklearn.model_selection import TimeSeriesSplit
from sklearn.model_selection import GridSearchCV


def create_model(lstm_1,lstm_2,dense):
    model = Sequential()
    model.add(LSTM(lstm_1, input_shape=(X_train.shape[1], X_train.shape[2]), return_sequences=True))
    model.add(LSTM(lstm_2))
    model.add(Dense(dense, activation='linear'))
    model.add(Dense(1, activation='linear'))
    model.compile(optimizer='adam', loss='mean_squared_error')
    return model

param_grid = param_grid = {
    'lstm_1': [32, 64, 128],
    'lstm_2': [32, 64, 128],
    'dense': [16, 32, 64]
}



tscv = TimeSeriesSplit(n_splits=5)

class TwoLayerFeedForward:
     def __call__():
         clf = Sequential()
         clf.add(Dense(9, activation='relu', input_dim=3))
         clf.add(Dense(9, activation='relu'))
         clf.add(Dense(3, activation='softmax'))
         clf.compile(loss='categorical_crossentropy', optimizer=SGD())
         return clf



grid = GridSearchCV(estimator=model, param_grid=param_grid, cv=tscv, n_jobs=-1)
grid_result = grid.fit(X_train, y_train)


print("The Best parameters: ", grid_result.best_params_)

best_model = grid_result.best_estimator_.model


NN_predictions = best_model.predict(X_test)

NN_mse = np.mean((y_test - NN_predictions.flatten()) ** 2)
NN_rmse = np.sqrt(NN_mse)
print("The Root Mean Squared Error for the Neural Network model is: ", NN_rmse)


NN_r_squared = r2_score(y_test, NN_predictions)

n = len(y_test)
p = 8
NN_adjusted_r_squared = 1 - (1 - NN_r_squared) * (n - 1) / (n - p - 1)

print('The OOS R-squared is: ', NN_r_squared)
print('The OOS Adjusted R-squared is: ', NN_adjusted_r_squared)


plt.figure(figsize=(12, 6))
plt.plot(y_test, label='Implied Volatility')
plt.plot(NN_predictions, label='NN Forecast')
plt.xlabel('Date Index')
plt.ylabel('Implied Volatility')
plt.title('Actual vs Predicted values')
plt.legend()
plt.show()



from tensorflow.keras.regularizers import L2


LSTM1 = grid_result.best_params_['lstm_1']
LSTM2 = grid_result.best_params_['lstm_2']
DENSE = grid_result.best_params_['dense']


model_REG = Sequential([
    LSTM(LSTM1, input_shape=(X_train.shape[1], X_train.shape[2]), return_sequences=True, kernel_regularizer= L2(0.001)),
    LSTM(LSTM2),
    Dense(DENSE, activation = 'linear'),
    Dense(K, activation = 'linear')
])


model_REG.compile(optimizer='adam', loss='mean_squared_error')


history_REG = model_REG.fit(X_train, y_train, epochs=N_epochs, batch_size=N_batch_size, validation_data=(X_test, y_test))


NN_REG_predictions = model_REG.predict(X_test)


NN_REG_mse = np.mean((y_test - NN_REG_predictions.flatten()) ** 2)
NN_REG_rmse = np.sqrt(NN_REG_mse)
print("The Root Mean Squared Error for the Neural Network model is: ", NN_REG_rmse)

NN_r_squared = r2_score(y_test, NN_REG_predictions)

n = len(y_test)
p = 8
NN_adjusted_r_squared = 1 - (1 - NN_r_squared) * (n - 1) / (n - p - 1)

print('The OOS R-squared is: ', NN_r_squared)
print('The OOS Adjusted R-squared is: ', NN_adjusted_r_squared)


plt.figure(figsize=(12, 6))
plt.plot(y_test, label='Implied Volatility')
plt.plot(NN_REG_predictions, label='NN Forecast')
plt.xlabel('Date Index')
plt.ylabel('Implied Volatility')
plt.title('Actual vs Predicted values')
plt.legend()
plt.show()

`

Answer 1

https://stackoverflow.com/a/77251441有一个答案。从keras迁移到scikeras，大体是相同的方法。但也存在一些差异。也许该链接包含您的答案。