Home Page

Case Study

Stochastic Model Selection via MC3

2025
  • R
  • Statistical Computing
  • Bayesian Methods

MC3 over add/remove-one neighborhoods; rcdd linearity checks; MH with neighbor-count correction; compared to greedy selections.

GitHub

Problem & Motivation:

Explore model space without exhaustive enumeration.

Data & Approach:

  • MC3 proposals; acceptance with neighborhood correction; chains diagnostics; rcdd validity filters.

Results:

  • Stochastic search recovered high-scoring subsets similar to greedy but with alternatives for correlated features.

Limitations:

Chain mixing depends on proposal tuning.