I've been learning about LSTMs and I commonly see them applied to the same type of task. For example, given a 1D list of values, predict which class they belong to. Or, given 8 values from 8 sensors, predict a state.

What I am trying to figure out, is if I have samples with shape (20, N), can I give each 20xN sample a binary classification using only an LSTM? I was hoping to treat each 20x1 column of the row as a timestep. Unfortunately I have only seen examples where each 20x1 column can be treated as a sequence. Is it possible for each timestep in an LSTM to have multiple input variables?

Thanks in advance.


Update: I have learned that this is possible. In keras, the parameters for the LSTM layer go in this order: batch, sequence_length, num_features

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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