A method to estimate the significance of coincident gravitational-wave observations from compact binary coalescence

TL;DR

This paper introduces a new statistical method to estimate the significance of gravitational-wave events from compact binary coalescence, effectively handling non-Gaussian noise without relying on time-shifting techniques.

Contribution

The proposed approach estimates false alarm probabilities using common search variables, capturing data non-Gaussianity without explicit time shifts, and detects events with multiple moderate significance signals.

Findings

Method successfully estimates false alarm probabilities in LIGO data.

Effective in identifying events with several moderate significance signals.

Validated with simulated and real gravitational-wave data.

Abstract

Coalescing compact binary systems consisting of neutron stars and/or black holes should be detectable with upcoming advanced gravitational-wave detectors such as LIGO, Virgo, GEO and {KAGRA}. Gravitational-wave experiments to date have been riddled with non-Gaussian, non-stationary noise that makes it challenging to ascertain the significance of an event. A popular method to estimate significance is to time shift the events collected between detectors in order to establish a false coincidence rate. Here we propose a method for estimating the false alarm probability of events using variables commonly available to search candidates that does not rely on explicitly time shifting the events while still capturing the non-Gaussianity of the data. We present a method for establishing a statistical detection of events in the case where several silver-plated (3--5$\sigma$) events exist but not necessarily any gold-plated ($>5\sigma$) events. We use LIGO data and a simulated, realistic, blind signal population to test our method.