Recent theories linked long gamma ray bursts (GRBs) to galaxies with rapid star formation or starburst; thus, we expect that long GRBs (LGRBs) are more frequent in midcluster galaxies where mergers and tidal interactions between gas-rich galaxies are more likely to occur. Yet there is no galaxy cluster known to be associated with LGRBs. We demonstrate that, based on deep, single-band Subaru Hyper Suprime-Cam observations, we may provide constraints on photometric redshifts of groups of galaxies. We compare three methods: cosmological approach, pseudoinverse matrix, and random forests to estimate galaxy and quasar redshifts. Comparing our results to spectroscopic redshifts of Sloan Digital Sky Survey's-detected extragalactic sources, random forests may provide the highest accuracy with as low as 17 percentage error. This is a powerful method to find clusters to place GRB host galaxies in their local environment.