EPIC - Easy Parameter Inference in Cosmology: The user's guide to the MCMC sampler

TL;DR

EPIC is a Python-based MCMC sampler designed for cosmological parameter inference, offering Bayesian analysis, model comparison, and advanced algorithms like Parallel Tempering to efficiently explore complex posterior distributions.

Contribution

This paper introduces EPIC, a new user-friendly MCMC sampler for cosmology that includes features like Parallel Tempering and customizable options, with detailed guidance for installation and usage.

Findings

Includes Parallel Tempering for multi-peaked distributions

Provides adaptive routines for improved efficiency

Offers comprehensive user instructions and customization options

Abstract

Easy Parameter Inference in Cosmology (EPIC) is another Markov Chain Monte Carlo (MCMC) sampler for Cosmology. It is implemented in Python and provides Bayesian parameter inference and model comparison based on the Bayesian evidence. The Parallel Tempering algorithm is included, which can help in the exploration of posterior distributions with two or more separated peaks. Adaptive routines for obtaining better efficiency with fine-tuned algorithms are being developed and will be available in future versions. In this user's guide, I give general instructions for installation and usage, including examples, and show how to modify the code in order to add new datasets and models.