Updates to change detection algorithms #161
Merged
Conversation
This file contains 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
…at caught bad SD solutions. (3) new logic for finding transition points that pools neghboring nonzero differences. (4) new logic for choosing baseline for making corrections. (5) new logic for selecting best weight during hyperparameter tuning (largest weight with in bound set as min holdout error + 2% of test error range)
…nal solve in SD formulation.
…first difference component).
…s optional kwarg for solving polished version of problem
bmeyers
had a problem deploying
to
test-deployment
GitHub Actions
Failure
Thistleman
approved these changes
Jul 16, 2024
pluflou
approved these changes
Jul 16, 2024
…o set it correctly in the test.
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Outdated suggestions cannot be applied.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.
Recent changes to our model of smooth, periodic functions led to issues with using QSS in nonconvex mode for change point detection. We are deprecating the use of QSS and standardizing the use of CLARABEL to solve the SD problems. In addition, we are no longer using the non-convex formulation for change point detection. Instead, we solve the L1-heuristic problem and polish the solution.
In this PR, we make changes to the two subroutines that use this pattern: time shift detection and capacity change detection. Also, we update the unit test fixtures to account for the new SD models, after manually checking the outputs in the jupyter notebooks which are in the test fixtures.