New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[BUG] temporal_train_test_split
with fh
does not split per unique time series
#4968
Comments
There are some known issues with I think we should have it dispatch to splitters, rather than having its own logic - that would solve the occasional odd behaviour, and the need to update it separately from the splitters. Should we make this part of a splitter refactor? |
are you saying that you are suggesting to refactor |
temporal_train_test_split
with fh
does not split per unique time seriestemporal_train_test_split
with fh
does not split per unique time series
almost - I think |
Ok got it, that's a good idea! I was about to disagree if the refactor meant converting |
Hm, perhaps the easiest is to write a new splitter, which has Then it's called under the hood of To get the split for This refactor would also make it easy to allow an arbitrary number of args similar to |
This can be a bug, or may be an expected behaviour as of now. In the latter case, I would want to submit this as a feature request that split by forecast horizon should work per series instead of on the entire dataset.
Reproducible code
Results
Version
The text was updated successfully, but these errors were encountered: