A Comparative Study of Machine-learning Methods for X-Ray Binary Classification

Abstract X-ray binaries (XRBs) consist of a compact object that accretes material from an orbiting secondary star. The most secure method we have for determining if the is black hole to determine its mass: This limited bright objects and requires substantial time-intensive spectroscopic monitoring. With new sources being discovered with different observatories, developing efficient, robust means classify becomes increasingly important. We compare three machine-learning classification methods (Bayesian Gaussian Processes (BGPs), K-Nearest Neighbors (KNN), Support Vector Machines) whether are neutron stars or holes (BHs) in XRB systems. Each uses spatial patterns exist between systems same type 3D color–color–intensity diagrams. used lightcurves extracted using 6 yr data MAXI/GSC 44 representative sources. find all highly accurate distinguishing pulsing nonpulsing (NPNS) 95% NPNS 100% pulsars accurately predicted. All high accuracy BHs (92%) but continue confuse subclass NPNS, called bursters, KNN doing best at only 50% predicting BHs. precision high, providing equivalent results over 5–10 independent runs. In future work, will suggest fourth dimension be incorporated mitigate confusion bursters. work paves way toward more efficiently distinguish BHs, pulsars.