将输入馈入keras顺序模型

Question

我正在与Keras进行时间序列预测。 我的输入数据中有10个时间步长，其形状为(2688, 10, 1) 2688，10，1 (2688, 10, 1)即train_x.shape和train_y.shape为（2688，10，1）。 尝试将其输入模型时出现以下错误。

ValueError: Error when checking target: expected activation_1 to have 2 dimensions, but got array with shape (2688, 10, 1)

我提供给第一lstm层的输入形状： input_shape=(1, time_steps) **

我可以正确地重塑train_y吗？

**

    time_steps=10
    train_x = np.reshape(train_x, (train_x.shape[0], train_x.shape[1], 1))
    train_y = np.reshape(train_y, (train_y.shape[0], train_y.shape[1], 1))

    # lstm model
    model = Sequential()
    model.add(LSTM(128, input_shape=(time_steps, 1), return_sequences=True))
    model.add(LSTM(64))
    model.add(Dense(1))
    model.add(Activation('linear'))

    model.compile(loss='mse', optimizer='adam')
    history = model.fit(train_x, train_y, epochs=10, validation_data=(test_x, 
    test_y), batch_size=64, verbose=1)

Answer 1

如果您期望形状（2688、10、1），则不能输入_shape =（1，time_steps）。 它应该input_shape =（time_steps，1）

Answer 2

train_x = np.reshape(train_x, (train_x.shape[0], train_x.shape[1], 1))
train_y = np.reshape(train_y, (train_y.shape[0], train_y.shape[1], 1)

我认为错误在这里，您应该给定shape [0]和shape [1]之间的时间步长。

train_x = np.reshape(train_x, (train_x.shape[0], 10,train_x.shape[1]))
train_y = np.reshape(train_y, (train_y.shape[0],10, train_y.shape[1]))

这里的值“ 10”表示时间步长！