LSTM模型搭建_LSTM神经网络

def LSTM_Classifier(train, trainLabel, test, testLabel, val_test, val_label, new_test=None):

    train, test = np.array(train), np.array(test)
    train, test = train.reshape(train.shape[0], 1, train.shape[1]), test.reshape(test.shape[0], 1, test.shape[1])
    val_test = np.array(val_test)
    val_test = val_test.reshape(val_test.shape[0], 1, val_test.shape[1])

    new_test = np.array(new_test)
    new_test = new_test.reshape(new_test.shape[0], 1, new_test.shape[1])


    trainLabel = np_utils.to_categorical(trainLabel)
    val_label = np_utils.to_categorical(val_label)

    # 单向LSTM
    model = Sequential()
    model.add(LSTM(360, activation='relu', input_shape=(train.shape[1], train.shape[2])))
    model.add(Dense(1024,activation='relu'))
    model.add(LeakyReLU(alpha=0.001))
    model.add(Dropout(0.4))
    model.add(Dense(2, activation='sigmoid'))


    # 双向LSTM
    # model = Sequential()
    # model.add(Bidirectional(LSTM(160,activation='relu', return_sequences=True), input_shape=(train.shape[1], train.shape[2])))
    # model.add(Bidirectional(LSTM(160, activation='relu')))
    #
    # model.add(Dense(2, activation='sigmoid'))

    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
    model.fit(train, trainLabel, batch_size=10, epochs=10, verbose=0, validation_data=(val_test, val_label), shuffle=True)

    pred_1 = model.predict_classes(test)
    pred_2 = model.predict_classes(new_test)


    return pred_1, pred_2