Estimating GRBs' cosmological distances

Abstract

Several thousand gamma-ray bursts have been observed but more than 500 events have known distances. Numerous papers have shown huge structures in the Universe based on GRBs. We want to examine the distribution of the GRBs' distances, to give an estimation on the distances of those GRBs which have no measured redshifts. The GRB catalogs contain more than 100 physical parameters among which there can be parameters dependent on distance. In this work we examined the distances of Swift GRBs with machine learning methods. For the regression of the distances we used both Random Forest and XGBoost regression algorithms. We found an 0.76 strength correlation between the regressed and measured redshift.