1
$\begingroup$

I'm trying to build a model to forecast multi-steps ahead time series data like stock market data.

And I found out that many precedents projects are done with many-to-many RNNs, not seq-to-seq.

Is there any reason for that? And what's the difference between those two algorithms?

$\endgroup$
1
$\begingroup$

As far as I can tell, the difference is more in the input/outputs than the structure of the network.

In a seq2seq model, you give the desired output, as a part of the input. And the output is the same as that part of the input, but shifted by one time step. The following chart is taken from here.

enter image description here

In contrast to the above architecture, a many to many model, does not include the output as a part of the input. As show here:

enter image description here

$\endgroup$

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.