Exploration of Regression Models for Homogenization of the Earthquake Magnitude Scales to Study Earthquakes in the Andaman-Nicobar Subduction Zone

The earthquake size is quantified by seismic centers all over the world in several magnitude scales, which requires to be unified into a preferred scale to infer flawless comparison of earthquake size. The problem has been addressed by numerous researchers and converted earthquake magnitudes mostly into moment magnitude from different magnitude scales as relation among them are overdue. In this study, Simple Linear Regression (SLR), Orthogonal Regression (OR), Gaussian Process Regression (GPR), and Support Vector Regression (SVR) have been employed to find out the best method to correlate different earthquake magnitude scales. The Andaman–Nicobar–Subduction Zone, a part of the Sumatra–Andaman–Subduction–Zone which is capable of producing earthquakes of great magnitudes, has been selected for this study. For this study, earthquake data has been extracted from the International Seismological Centre (ISC) earthquake catalog. The study indicates that moment magnitude can be predicted more accurately using multiple scales than single scale. It has also been found that, the GPR model predicts higher earthquake magnitude values better than other methods. Hence, for conversion of earthquake magnitude into standard scale, use of multiple scales and the GPR method can be beneficial.