Testing the normality of the gravitational wave data with a low cost recursive estimate of the kurtosis

TL;DR

This paper introduces a low-cost recursive method to monitor the normality of gravitational wave detector data by estimating kurtosis, validated through simulations.

Contribution

It presents a novel recursive approach for real-time kurtosis estimation to assess data normality in gravitational wave detection.

Findings

Recursive kurtosis estimation is effective for normality monitoring.

The method is computationally efficient and suitable for real-time applications.

Validation with simulated noise demonstrates its reliability.

Abstract

We propose a monitoring indicator of the normality of the output of a gravitational wave detector. This indicator is based on the estimation of the kurtosis (i.e., the 4th order statistical moment normalized by the variance squared) of the data selected in a time sliding window. We show how a low cost (because recursive) implementation of such estimation is possible and we illustrate the validity of the presented approach with a few examples using simulated random noises.