ValueError：对于多元时间序列数据，输入 arrays 应具有与目标 arrays 相同数量的样本

Question

我想创建一个具有多个 output 的多元 LSTM model，但我面临 X 和 y 数据的输入形状问题。 找不到任何解决方案。

我担心的一个问题是，如果我们使用多元时间序列，我们是否必须在 output 上使用 3 个神经元？

实际数据：
df.head()

               Tem.        Hum        Presure
2019-01-01  16.808511   66.085106   1018.042553
2019-01-02  19.437500   63.270833   1015.020833
2019-01-03  20.416667   68.312500   1014.104167
2019-01-04  27.500000   77.395833   1008.458333
2019-01-05  31.854167   73.708333   1004.708333

进行数据预处理后

X = 

    array([[[0.        , 0.22182641, 0.9932125 ]],
    
           [[0.17473411, 0.129819  , 0.86032281]],
    
           [[0.23981381, 0.29464689, 0.82000949]],
    
           ...,
    
           [[0.31314616, 0.53057328, 0.86911437]],
    
           [[0.30827762, 0.55054247, 0.87090296]],
    
           [[0.30306702, 0.55759395, 0.87138509]]])

y =
array([[0.29757451, 0.56409832, 0.87433447]])

X.shape = (349, 1, 3)

y.shape = (1, 3)

model = Sequential()
model.add(LSTM(200, activation='relu', return_sequences=True,input_shape=(X.shape[1], X.shape[2])))
model.add(LSTM(50,activation='relu',return_sequences=False))
model.add((Dense(3)))
model.compile(optimizer='adam', loss='mse')

early_stop = EarlyStopping(patience=5, restore_best_weights=True)

model.summary()

 Model: "sequential_18"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
lstm_28 (LSTM)               (None, 1, 200)            163200    
_________________________________________________________________
lstm_29 (LSTM)               (None, 50)                50200     
_________________________________________________________________
dense_16 (Dense)             (None, 3)                 153       
=================================================================
Total params: 213,553
Trainable params: 213,553
Non-trainable params: 0
_________________________________________________________________
"

history=model.fit(X, y, epochs=50, verbose=2, shuffle=False, callbacks=[early_stop],use_multiprocessing=True,)

谢谢 ！