OpenSkillIndex← back to search
claude-codescience

Babysitter emcee-mcmc-sampler

emcee MCMC skill for Bayesian parameter estimation and posterior sampling in physics applications

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 "$T/library/specializations/domains/science/physics/skills/emcee-mcmc-sampler" ~/.claude/skills/a5c-ai-babysitter-emcee-mcmc-sampler && rm -rf "$T"
manifest: library/specializations/domains/science/physics/skills/emcee-mcmc-sampler/SKILL.md
tags
#bayesian-inference#mcmc-sampling#physics-analysis#posterior-sampling#parameter-estimation
source content

emcee MCMC Sampler

Purpose

Provides expert guidance on emcee for Bayesian parameter estimation in physics, including ensemble sampling and convergence diagnostics.

Capabilities

  • Affine-invariant ensemble sampling
  • Parallel tempering support
  • Autocorrelation analysis
  • Convergence diagnostics
  • Prior/likelihood specification
  • Chain visualization

Usage Guidelines

  1. Model Setup: Define log-probability function
  2. Initialization: Initialize walkers appropriately
  3. Sampling: Run ensemble sampler
  4. Convergence: Check autocorrelation and convergence
  5. Analysis: Extract posterior distributions

Tools/Libraries

  • emcee
  • corner
  • arviz