Question

我有9000个序列，每个序列的长度为200，只有一个功能。

#data.shape= (9000,200,1)

我想基于长度为190的输入序列来预测长度为200的序列。 X是长度为190的输入序列，Y是长度为200的输出序列。

X = np.delete(data,slice(50,60),1)  # shape of X = (9000,190,1)
Y = data.copy() # shape of Y = (9000,200,1)

我的问题基于教程Encoder-Decoder Model for Sequence-to-Sequence Prediction

以及现有的stackoverflow问题 seq2seq prediction for time series

# returns train, inference_encoder and inference_decoder models
def define_models(n_input, n_output, n_units):
    # define training encoder
    encoder_inputs = Input(shape=(None, n_input))
    ## First LSTM Layer (Encoder layer)
    encoder = LSTM(n_units, return_state=True)
    encoder_outputs, state_h, state_c = encoder(encoder_inputs)
    encoder_states = [state_h, state_c]

    # define training decoder
    decoder_inputs = Input(shape=(None, n_output))
    ## Second LSTM Layer (Decoder layer)
    decoder_lstm = LSTM(n_units, return_sequences=True, return_state=True) # Decoder returns sequence, while encoder do not
    decoder_outputs, _, _ = decoder_lstm(decoder_inputs, initial_state=encoder_states) # initial state of decoder is set to encoder states
    ## Dense Layer of decoder
    decoder_dense = Dense(n_output, activation='softmax') 
    decoder_outputs = decoder_dense(decoder_outputs)
    model = Model([encoder_inputs, decoder_inputs], decoder_outputs) # this is trained model containing both encoder and decoder

    # define inference encoder
    encoder_model = Model(encoder_inputs, encoder_states)

    # define inference decoder
    decoder_state_input_h = Input(shape=(n_units,))
    decoder_state_input_c = Input(shape=(n_units,))
    decoder_states_inputs = [decoder_state_input_h, decoder_state_input_c]
    decoder_outputs, state_h, state_c = decoder_lstm(decoder_inputs, initial_state=decoder_states_inputs)
    decoder_states = [state_h, state_c]
    decoder_outputs = decoder_dense(decoder_outputs)
    decoder_model = Model([decoder_inputs] + decoder_states_inputs, [decoder_outputs] + decoder_states)
    # return all models
    return model, encoder_model, decoder_model

train, infenc, infdec = define_models(1, 1, 128)
train.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['acc'])

在本教程中，作者使用两个输入序列来拟合模型，例如

train.fit([X1, X2], y, epochs=1)

我的问题是我该如何解决我的问题（当我只有一个功能时，如何创建数组X2（上下文向量））？ X2是根据教程的目标输出的移位序列。

如何仅使用一项功能为序列预测序列建模数据

0 个答案: