0
$\begingroup$

Let's say one has many time series for which one wants to build a predictive model (based on LSTM). Which of the following cases would be more optimal and why?

1) building one model for all the time series in which each time series has a categorical label which will be included in the model by one hot encoding.

2) building individual models for each time series.

$\endgroup$
  • $\begingroup$ Will be good if you provide sample data set. $\endgroup$ – vipin bansal Jul 29 at 5:02
  • $\begingroup$ Let's say there are around 10,000 time series. $\endgroup$ – physics_2015 Jul 29 at 5:21
  • $\begingroup$ You need a 10,000 time series output? $\endgroup$ – Leevo Jul 29 at 7:03
  • $\begingroup$ yes. An output for each time series. $\endgroup$ – physics_2015 Jul 29 at 17:02

Your Answer

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

Browse other questions tagged or ask your own question.