staar: multi-binary, multi-kinship, and random.slope nulls#133
Merged
staar: multi-binary, multi-kinship, and random.slope nulls#133
Conversation
Port MultiNull::fit_binary (joint k-binary, lift of single-trait logistic + MultiSTAAR Kronecker; GMMAT rejects multi-binomial so no upstream), multi_kinship::fit_multi_kinship (port of GMMAT glmmkin.multi.ai), and kinship::fit_reml_random_slope (port of glmmkin random.slope expansion with PSD-box step-halving). k=1 parity bit-for-bit vs single-trait on all three paths. Closes #111, #80, #101.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Ports the three remaining null-fit extensions to finish v0.3.0.
MultiNull::fit_binaryfits k binary traits jointly (per-trait logistic IRLS + cross-trait residual correlation + shared-W̄ Fisher-information projection). GMMAT rejects multi-binomial atglmmkin.R:43so this is a lift combining our single-trait logistic with the MultiSTAAR Kronecker pattern, not a 1:1 port.multi_kinship::fit_multi_kinshipis the 1:1 port ofglmmkin.multi.ai(glmmkin.R:424-536) — expands the kinship list viaE_{j,m} ⊗ Vblocks, stacksvec(Y)andI_k ⊗ X, runs AI-REML, reshapes theta back to per-group/per-kinship(k × k).kinship::fit_reml_random_slopeports therandom.slopeexpansion (glmmkin.R:174-183) with PSD-box step-halving (:375-376).k=1 parity is bit-for-bit against the single-trait paths on all three. 221
staar::tests green, invariance golden unchanged, clippy clean.Dense-only for multi-kinship; sparse Σ on the stacked nk × nk system is the natural follow-up. R ground-truth fixtures for k>1 multi-kinship and for random.slope are not in this PR — the k=1 reductions give strong confidence but not full numerical parity. Binary + kinship in multi-trait and binary + random.slope are rejected with clear messages (GMMAT doesn't implement either).
Closes #111
Closes #80
Closes #101