Particle identification in KM3NeT/ORCA

• Published in: Proceedings of 38th International Cosmic Ray Conference — PoS(ICRC2023)
• Main Authors: Cerisy, L, Pedrajas, A.L., Lastoria, C, Perrin-Terrin, M, Brunner, J, Dabhi, V
• Other Authors: Centre de Physique des Particules de Marseille (CPPM), Aix Marseille Université (AMU)-Institut National de Physique Nucléaire et de Physique des Particules du CNRS (IN2P3)-Centre National de la Recherche Scientifique (CNRS), Universitat de València (UV), KM3NeT
• Format: Conference Object
• Language: English
• Published: HAL CCSD 2023
• Subjects: machine learning; neutral current; flavor; Grid computing; particle identification; electron: interaction; atmosphere; neutrino: oscillation; charged current; KM3NeT; muon; [PHYS.ASTR]Physics [physics]/Astrophysics [astro-ph]; Orca
• Online Access: https://hal.science/hal-04186231; https://hal.science/hal-04186231/document; https://hal.science/hal-04186231/file/Particle%20identification%20in%20KM3NeT%20ORCA.pdf; https://doi.org/10.22323/1.444.1191

International audience One of the main goals of KM3NeT/ORCA is to measure atmospheric neutrino oscillation parameters with competitive precision. To achieve this goal, good discrimination between track-like and shower-like events is necessary, with particular focus on the measurement of the tau neutrino normalisation. The track-like signal is mainly carried by muon neutrinos from charged current interactions, while the shower-like signal comes from charged current interactions of electron and tau neutrinos, and neutral current interactions of all flavours. A Random Grid Search algorithm is optimised to separate these channels and its performance is compared with machine learning methods using boosted decision trees. This contribution will report on the technical aspects of the algorithm and the performance of the particle identification with data recorded in 2020 and 2021 using an early six-lines configuration of the ORCA detector (ORCA6).