feat: Gibbs variable selection + single-chain default for MCMC#74
Merged
feat: Gibbs variable selection + single-chain default for MCMC#74
Conversation
…ing) New MCMC mode: method='gibbs' (default) vs method='sbs' (original) Gibbs variable selection samples feature inclusion/exclusion within the MCMC iterations, with a sparsity prior P(k) = Binomial(k; n, p0). This replaces the O(n_features) SBS loop with a single MCMC run. New params: - mcmc.method: 'gibbs' (default) or 'sbs' - mcmc.p0: prior inclusion probability (default 0.1) Adaptive BTR threshold: selects features with P(active) > 0.5, falls back to top-k by posterior if too few qualify. Refs #70, #73
…aram Parallel chains with short n_iter don't converge — each chain burns n_burn iterations independently, leaving too few effective samples. Default to single chain; users can set n_chains > 1 when n_iter is large enough.
This was referenced Mar 25, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
method=gibbs(default) — joint feature+coefficient sampling in a single MCMC runmethod(gibbs/sbs),p0(prior inclusion prob),n_chainsBenchmark (Qin2014)
Gibbs produces sparser models with more FBM diversity.
Test plan
Refs #70, #73