OpenSkillIndex← back to search
claude-codescience

Babysitter mcmc-diagnostics

MCMC convergence diagnostics and analysis

install
source · Clone the upstream repo
git clone https://github.com/a5c-ai/babysitter
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/a5c-ai/babysitter "$T" && mkdir -p ~/.claude/skills && cp -r "$T/library/specializations/domains/science/mathematics/skills/mcmc-diagnostics" ~/.claude/skills/a5c-ai-babysitter-mcmc-diagnostics && rm -rf "$T"
manifest: library/specializations/domains/science/mathematics/skills/mcmc-diagnostics/SKILL.md
tags
#mcmc-analysis#bayesian-inference#convergence-check#statistical-diagnostic#rhat-calculation#ess-analysis
source content

MCMC Diagnostics

Purpose

Provides MCMC convergence diagnostics and analysis capabilities for validating Bayesian inference results.

Capabilities

  • Rhat (potential scale reduction) computation
  • Effective sample size (ESS) calculation
  • Trace plot generation
  • Autocorrelation analysis
  • Divergence detection
  • Energy diagnostic (E-BFMI)

Usage Guidelines

  1. Convergence Check: Verify Rhat < 1.01 for all parameters
  2. Sample Quality: Ensure ESS is sufficient for inference
  3. Visual Inspection: Review trace plots for mixing
  4. Divergences: Address divergent transitions

Tools/Libraries

  • ArviZ
  • CODA
  • MCMCpack