Characteristic sky background features around galaxy mergers

Abstract

In the context of finding galaxy mergers in large-scale surveys, we applied machine-learning algorithms that made use of flux measurements instead of using images (as is the current standard). By training multiple NNs using the Sloan Digital Sky Survey class-balanced data set of mergers and non-mergers, we found that sky-background error parameters could provide a validation accuracy of 92.64 ± 0.15% and a training accuracy of 92.36 ± 0.21%. Moreover, analyzing the NN identifications led us to find that a simple decision diagram using the sky error for two flux filters was enough to gain a 91.59% accuracy. By understanding how the galaxies vary along the diagram and trying to parametrize the methodology in the deeper images of the Hyper Suprime Cam, we are currently trying to define and generalize this sky error-based methodology.