Machine learning based TOF charged particle identification at BM@N detector of NICA collider

Abstract

Abstract In the article results of charged particles identification for BM@N experiment being performed at NICA acceleration complex of Joint Institute for Nuclear Research are presented. A standard neural network-based technique of constructing a classificator is applied to the data sets obtained both from modelling of a realistic experimental setup and three synthetic data sets. The carried-out analysis demonstrates that the estimated data accuracy is insufficient to make a clear distinction between electrons, muons and pions, and also between -particles and deutrons. The problem could be solved by using an extra data from the detector or by improving the accuracy of the experimental data by two orders of magnitude.