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
We want to uncouple DeadlineMovingAverageModel from PerSegmentModelMixin. Doing this, we can make DeadlineMovingAverageModel predict on any new segments because it doesn't require fit to work.
It also can lead to better performance.
Proposal
Implement DeadlineMovingAverageModel without using PerSegmentModelMixin and NonPredictionIntervalContextRequiredModelMixin. It should be inherited only from NonPredictionIntervalContextRequiredAbstractModel.
Method get_model should return self.
Try to vectorize forecasting logic.
Test cases
Make sure current tests pass.
Adapt tests on _DeadlineMovingAverageModel.
Fix test_inference tests to work on new segments.
Additional context
No response
The text was updated successfully, but these errors were encountered:
Mr-Geekman
changed the title
Uncouple DeadlineMovingAverageModel from PerSegmentModelMixin
Decouple DeadlineMovingAverageModel from PerSegmentModelMixinFeb 15, 2023
🚀 Feature Request
We want to uncouple
DeadlineMovingAverageModel
fromPerSegmentModelMixin
. Doing this, we can makeDeadlineMovingAverageModel
predict on any new segments because it doesn't require fit to work.It also can lead to better performance.
Proposal
DeadlineMovingAverageModel
without usingPerSegmentModelMixin
andNonPredictionIntervalContextRequiredModelMixin
. It should be inherited only fromNonPredictionIntervalContextRequiredAbstractModel
.get_model
should returnself
.Test cases
_DeadlineMovingAverageModel
.test_inference
tests to work on new segments.Additional context
No response
The text was updated successfully, but these errors were encountered: