I have a list of stock price sequences with 20 timesteps each. That's a 2D array of shape (total_seq, 20). I can reshape it into (total_seq, 20, 1) for concatenation to other features.

I also have news title with 10 words for each timestep. So I have 3D array of shape (total_seq, 20, 10) of the news' tokens from Tokenizer.texts_to_sequences() and sequence.pad_sequences().

I want to concatenate the news embedding to the stock price and make predictions.

My idea is that the news embedding should return tensor of shape (total_seq, 20, embed_size) so that I can concatenate it with the stock price of shape (total_seq, 20, 1) then connect it to LSTM layers.

To do that, I should convert news embedding of shape (total_seq, 20, 10) to (total_seq, 20, 10, embed_size) by using Embedding() function.

But in Keras, the Embedding() function takes a 2D tensor instead of 3D tensor. How do I get around with this problem?

Assume that Embedding() accepts 3D tensor, then after I get 4D tensor as output, I would remove the 3rd dimension by using LSTM to return last word's embedding only, so output of shape (total_seq, 20, 10, embed_size) would be converted to (total_seq, 20, embed_size)

But I would encounter another problem again, LSTM accepts 3D tensor not 4D so

How do I get around with Embedding and LSTM not accepting my inputs?

  • 1
    $\begingroup$ It's a bit funky how to do this, but I'm writing an example for you! $\endgroup$ Commented Aug 11, 2017 at 13:09
  • $\begingroup$ @JanvanderVegt Please check my question again. I added more details! Thanks. $\endgroup$
    – offchan
    Commented Aug 11, 2017 at 13:26
  • 1
    $\begingroup$ Instead of what you wanted I got to an output after the embeddings of (total_seq, 20, 10 * embed_size + 1) by concatenating the embeddings for each of the words and the stock price, would that solve your problem? $\endgroup$ Commented Aug 11, 2017 at 13:42
  • $\begingroup$ It would be better if I can somehow summarize the news into one small embedding for each time step. $\endgroup$
    – offchan
    Commented Aug 11, 2017 at 17:32
  • 2
    $\begingroup$ That's easy, add a TimeDistrbuted(Dense(dim)) that takes the concatenated words. You could use a convolution or RNN here on the words but with only 10 words I think Dense might be better. I updated my example, I hope this is what you were looking for. $\endgroup$ Commented Aug 11, 2017 at 19:16

1 Answer 1


I'm not entirely sure if this is the cleanest solution but I stitched everything together. Each of the 10 word positions get their own input but that shouldn't be too much of a problem. The idea is to make an Embedding layer and use it multiple times. First we will generate some data:

n_samples = 1000
time_series_length = 50
news_words = 10
news_embedding_dim = 16
word_cardinality = 50

x_time_series = np.random.rand(n_samples, time_series_length, 1)
x_news_words = np.random.choice(np.arange(50), replace=True, size=(n_samples, time_series_length, news_words))
x_news_words = [x_news_words[:, :, i] for i in range(news_words)]
y = np.random.randint(2, size=(n_samples))

Now we will define the layers:

## Input of normal time series
time_series_input = Input(shape=(50, 1, ), name='time_series')

## For every word we have it's own input
news_word_inputs = [Input(shape=(50, ), name='news_word_' + str(i + 1)) for i in range(news_words)]

## Shared embedding layer
news_word_embedding = Embedding(word_cardinality, news_embedding_dim, input_length=time_series_length)

## Repeat this for every word position
news_words_embeddings = [news_word_embedding(inp) for inp in news_word_inputs]

## Concatenate the time series input and the embedding outputs
concatenated_inputs = concatenate([time_series_input] + news_words_embeddings, axis=-1)

## Feed into LSTM
lstm = LSTM(16)(concatenated_inputs)

## Output, in this case single classification
output = Dense(1, activation='sigmoid')(lstm)

After compiling the model we can just fit it like this:

model.fit([x_time_series] + x_news_words, y)


After what you mentioned in the comments, you can add a dense layer that summarizes the news, and adds that to your time series (stock prices):

## Summarize the news:
news_words_concat = concatenate(news_words_embeddings, axis=-1)
news_words_transformation = TimeDistributed(Dense(combined_news_embedding))(news_words_concat)

## New concat
concatenated_inputs = concatenate([time_series_input, news_words_transformation], axis=-1)
  • $\begingroup$ works like a charm $\endgroup$
    – offchan
    Commented Aug 13, 2017 at 9:51
  • $\begingroup$ Perfect :) good luck! $\endgroup$ Commented Aug 13, 2017 at 10:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.