A Bayesian Surrogate Model for Rapid Time Series Analysis and Application to Exoplanet Observations

TL;DR

This paper introduces a Bayesian surrogate model designed for rapid analysis of periodic and quasi-periodic time series data, with applications demonstrated on exoplanet observations, highlighting computational efficiency and practical challenges.

Contribution

The paper presents a novel Bayesian surrogate model with efficient computation for time series analysis, specifically tailored for exoplanet data, including model comparison capabilities.

Findings

Effective analysis of simulated exoplanet data

Challenges identified in real-world data application

Importance of observation spacing and prior selection

Abstract

We present a Bayesian surrogate model for the analysis of periodic or quasi-periodic time series data. We describe a computationally efficient implementation that enables Bayesian model comparison. We apply this model to simulated and real exoplanet observations. We discuss the results and demonstrate some of the challenges for applying our surrogate model to realistic exoplanet data sets. In particular, we find that analyses of real world data should pay careful attention to the effects of uneven spacing of observations and the choice of prior for the "jitter" parameter.