Unsupervised Deep Learning Method for Clustering KAGRA O3GK Transient Noise Data

Abstract

The advent of ground-based gravitational wave detectors has significantly improved the detection of faint gravitational wave signals. However, these detectors are affected by various types of transient noise, known as glitches, which can mimic true gravitational wave signals and limit detector sensitivity. Classifying glitches according to their time-frequency characteristics not only facilitates a deeper understanding of their origins, but also supports their mitigation on data to maintain data…