Probabilistic Cross-Identification of Cosmic Events

TL;DR

This paper introduces a Bayesian method for cross-identifying cosmic events like supernovae across different observations, leveraging lightcurve data and probabilistic modeling to improve accuracy and efficiency.

Contribution

It presents a novel Bayesian hypothesis testing framework for associating cosmic events across independent datasets, incorporating lightcurve analysis and Gaussian approximations.

Findings

Bayesian approach effectively matches cosmic events in space and time.

Gaussian approximation of lightcurves simplifies the matching process.

Model-dependent associations increase confidence in event identification.

Abstract

We discuss a novel approach to identifying cosmic events in separate and independent observations. In our focus are the true events, such as supernova explosions, that happen once, hence, whose measurements are not repeatable. Their classification and analysis have to make the best use of all the available data. Bayesian hypothesis testing is used to associate streams of events in space and time. Probabilities are assigned to the matches by studying their rates of occurrence. A case study of Type Ia supernovae illustrates how to use lightcurves in the cross-identification process. Constraints from realistic lightcurves happen to be well-approximated by Gaussians in time, which makes the matching process very efficient. Model-dependent associations are computationally more demanding but can further boost our confidence.