Classification algorithms applied to structure formation simulations

Throughout cosmological simulations, the properties of matter density field in initial conditions have a decisive impact on features structures formed today. In this paper we use random-forest classification algorithm to infer whether or not dark particles, traced back conditions, would end up halos whose masses are above some threshold. This problem might be posed as binary task, where mapped into labels provided by halo finder program. Our results show that random forests effective tools predict output simulations without running full process. These techniques used future decrease computational time and explore more efficiently effect different matter/dark energy candidates formation structures.