Prune failed Matching trials instead of scoring the training mean - #215
Open
vahid-ahmadi wants to merge 2 commits into
Open
vahid-ahmadi wants to merge 2 commits into
vahid-ahmadi wants to merge 2 commits into
Conversation
Both tuning handlers caught every exception and substituted the training mean, with no log and no counter, and that score went straight into the Optuna objective. A mean-predictor is not a neutral score - on a low-signal target it can beat a genuine matching fit on quantile loss - so a parameter set under which matching always failed could be selected as best and reported as the winning method. Both now log the exception and raise TrialPruned. The predict path keeps its NaN fill, which is the right behaviour there, but now reports the total number of unmatched records rather than leaving silent NaN blocks. Fixes #210
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #210.
Matching tuning prunes failed donor-matching trials so fallback training means cannot win model selection. A failure in a later chunk invalidates the whole trial. If all trials fail, tuning raises an explicit Matching error.
Chunked predictions retain successful rows and report failed rows as missing values.
n_failed_recordsstarts at zero and reflects the latest returned prediction, including small predictions after a previous chunk failure. The Markdown examples also pass the repository formatter.Validation: eight deterministic backend regressions exercise real Optuna pruning, all-pruned behavior, partial output, index preservation and failure-count resets. They first produced 5 failures and 3 passing controls; the affected suite then passed 61 tests with one optional R module skipped locally. Final real-R CI results are recorded below.
Final commit:
33501b6547945ecb37c4e8e273193f2988491cf2. All eight checks passed. Python 3.14 CI passed 361 tests with 2 skipped, including all 13 real R/StatMatch Matching tests; the pipeline example passed. Python 3.12 smoke tests, documentation, lint, changelog and deployment checks also passed. The two skipped modules require optional MDN dependencies. Coverage XML was generated; the existing Codecov upload step reported its nonblocking protected-branch token error.The exact final heads of #214, #215, #216 and #217 also merge cleanly together. Their combined suite passed 350 tests with 3 optional skips, and the combined source distribution, wheel, import and numerical smoke checks passed. Merge these before #219, then rebase #219 while retaining its broader correctness fixes.