Picking-up Local GRB Candidates Based on Their Host Galaxies

Abstract Rapid identification of candidates of high-value gamma-ray bursts (GRBs), including both high- z and local events, is crucial for outlining subsequent observational strategy. In this paper, we present a model that enables an on-duty astronomer to rapidly identify candidates of local GRBs prior to spectroscopy, provided that these events have been localized at an arcsecond's precision. After taking into account the mass distribution of the host galaxies of GRBs, the model calculates the two-dimensional cross-match probabilities between a localized GRB and its surrounding nearby galaxies, and then returns the best match with the highest probability. The model is evaluated not only by the observed GRB sample with redshifts up to z = 4, but also through the simulated GRB samples. By using the recently published GLADE+ galaxies catalog with a completeness of 95% up to 500 Mpc, along with the NED-LVS catalog, the Precision and Recall of the model are determined to be 0.23–0.33 and 0.75, respectively, at the best performance. A dedicated web service, which will be integrated into the SVOM Science User Support System, has been developed to deploy the model.