Separation of $e^+e^-\to e^+e^-$ and $e^+e^-\to\pi^+\pi^-$ events using SND detector calorimeter

TL;DR

This paper presents a machine learning-based method to distinguish between electron-positron and pion-pair events in collider experiments, achieving high identification efficiency in the 0.5 to 1 GeV energy range.

Contribution

The study introduces a novel machine learning technique for event discrimination in the SND detector, significantly improving identification efficiency over previous methods.

Findings

Identification efficiency for $e^+e^- o e^+e^-$ and $e^+e^- o \pi^+\pi^-$ events reached 99.3-99.8%.

Effective discrimination in the energy range 0.5 to 1 GeV.

Demonstrated the applicability of machine learning in particle event classification.

Abstract

The technique of discrimination of the $e^+e^-\to e^+e^-$ and $e^+e^-\to \pi^+\pi^-$ events in energy range $0.5 < \sqrt{s} < 1$ GeV by energy deposition in the calorimeter of SND detector was developed by applying machine learning method. Identification efficiency for $e^+e^-\to e^+e^-$ and $e^+e^-\to \pi^+\pi^-$ events in the range from 99.3 to 99.8 % has been achived.