Binary Black Hole Automated Identification by Agglomerative Clustering based on Gravitational Waves

Abstract

The General Theory of Relativity, proposed by Albert Einstein theoretically predicted that very large accelerating mass creates ripples in spacetime which is the strongest for merging binary black hole system and the ripples can travel billions of light-years and these ripples are called Gravitational Waves. By the time these waves reach Earth, they become very faint and can’t be detected with regular methods. For this, LIGO has created specialized detectors based on the laser interference principle to detect strains caused by gravitational waves in e-19 scale. GW190521 is a gravitational wave event recorded on 21 May 2019 at 03:02:29 UTC and caused by the merger of two black holes of 85M© and 66 M© whose progenitor was the largest ever recorded. Throughout literature, very few amounts of autonomous black hole identification models have been made because of limited data availability. This experiment proposes methods for autonomous identification of black holes by using an unsupervised machine learning algorithm called Agglomerative Clustering with very little data to train which can adapt quickly to gravitational wave events. The model could be easily deployed near laser interferometric observatories for autonomous black hole identification with minimal effort.