Fix ordinal cloglog probability transform for K>=4 categories#382
Merged
Conversation
The sequential ordinal cloglog model uses P(Y=k) = prod_{j<k} S_j * (1-S_k)
where S_k = exp(-exp(gamma_k + f)). The old code used S_{k-1}*(1-S_k) for
intermediate categories (k>=2), which only accidentally gave the correct answer
for K=3 (where c_1 = gamma_1 always), but silently produced wrong class
probabilities for K>=4.
Replace with a running-product loop that tracks cumulative survival correctly
across all K. Add K=4 regression tests in both R and Python that verify the
predict output matches the manual formula and sums to 1.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The sequential ordinal cloglog model uses$P(Y=k) = \prod_{j<k} S_j * (1-S_k)$ where $S_k = \exp(-exp(\gamma_k + f(X)))$ . The old code used $S_{k-1}*(1-S_k)$ for intermediate categories ($c_1 = \gamma_1$ always), but silently produced wrong class probabilities for
k>=2), which only accidentally gave the correct answer forK=3(whereK>=4.This PR replaces that faulty code with a running-product loop that tracks cumulative survival correctly across all K. We also add explicit unit tests for the K > 3 case in both R and Python that verify the predict output matches the manual formula and sums to 1.