You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello,
I have two problems with my forecast with DeepAr based on mxnet.
So first, the forecast intervals are way too small, I asked about this before and got as answer that I should use the validation sampler. I did that, but unfortunately I couldn't see any particular improvement in the prediction interval.
The other problem is really very strange. I use the same data and the same features. If I re-train the model, that is create a new model with the same data and same hyperparameters etc. and then run the prediction and evaluation, I always get completely different results. The NRMSE varies from 2 to 0.5 depending on the run. is this normal? what can I do about the problems?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hello,
I have two problems with my forecast with DeepAr based on mxnet.
So first, the forecast intervals are way too small, I asked about this before and got as answer that I should use the validation sampler. I did that, but unfortunately I couldn't see any particular improvement in the prediction interval.
The other problem is really very strange. I use the same data and the same features. If I re-train the model, that is create a new model with the same data and same hyperparameters etc. and then run the prediction and evaluation, I always get completely different results. The NRMSE varies from 2 to 0.5 depending on the run. is this normal? what can I do about the problems?
Beta Was this translation helpful? Give feedback.
All reactions