[ML] Fix regression in QA data set #489
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
edsavage
approved these changes
Jun 3, 2019
| @@ -1619,6 +1619,8 @@ void CUnivariateTimeSeriesModel::reinitializeStateGivenNewComponent( | |||
| // re-weight so that the total sample weight corresponds to the sample | |||
| // weight the model receives from a fixed (shortish) time interval. | |||
|
|
|||
| m_ResidualModel->removeModels( | |||
There was a problem hiding this comment.
Perhaps add a comment to explain why we're removing Poisson from model selection in this case?
There was a problem hiding this comment.
I added this comment
tveasey
added a commit
to tveasey/ml-cpp-1
that referenced
this pull request
Jun 10, 2019
tveasey
added a commit
that referenced
this pull request
Jun 10, 2019
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
One of our QA sets has shown a regression associated with #468. This was triggered by selecting our Poisson process model in conjunction with a periodic event rate. Including the Poisson process at all in such cases is low value and since we don't properly deal with the change in variance associated with rate changes in this case it has undesirable side effects. This change simply removes it from model selection if we start modelling seasonality (we can revisit this if we do any work to better deal with the variance).
I also increased the parameter introduced in #468 (this means we revert slightly more to our old behaviour). Finally, I've rejigged reinitialisation of the multi-bucket model decay rate, which should have been matching the single bucket model.