Using real static features results in non-decreasing loss? #2535
Unanswered
RashidBakirov
asked this question in
Q&A
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I am forecasting for several "target" categories 12 months ahead, while for several others "known" categories the values for these months ahead are known.
I have tried to use these known values as real static features, however, doing so results in a non-decreasing loss, which doesn't happen otherwise. I would imagine that if these features would not be useful, including them should not be very detrimental at least. Is there something that I need to consider?
On a related note, is this the preferred way of including "future" data i training? Another way I have tried was just including all available data in the training dataset, where "known" categories had data for one year longer; this seemed to be better than not including this additional year.
Beta Was this translation helpful? Give feedback.
All reactions