Automatic Bayesian inference for LISA data analysis strategies

TL;DR

This paper introduces an automatic Bayesian inference method using Reversible Jump MCMC for analyzing complex LISA gravitational wave data, effectively identifying the number and parameters of overlapping signals.

Contribution

It presents a novel automatic Reversible Jump MCMC algorithm for Bayesian inference in LISA data analysis, capable of handling unknown numbers of signals.

Findings

Successfully applied to simulated LISA data with overlapping signals

Effectively identified signals from white dwarf binaries and EMRIs

Demonstrated viability for future LISA data analysis challenges

Abstract

We demonstrate the use of automatic Bayesian inference for the analysis of LISA data sets. In particular we describe a new automatic Reversible Jump Markov Chain Monte Carlo method to evaluate the posterior probability density functions of the a priori unknown number of parameters that describe the gravitational wave signals present in the data. We apply the algorithm to a simulated LISA data set containing overlapping signals from white dwarf binary systems (DWD) and to a separate data set containing a signal from an extreme mass ratio inspiral (EMRI). We demonstrate that the approach works well in both cases and can be regarded as a viable approach to tackle LISA data analysis challenges.