Speaker
Description
The Zero-Degree Calorimeter(ZDC) is an important sub-detector component in the ePIC detector of the EIC. The detector is situated very close to the beam interaction point and consists of ECAL and HCAL to capture the neutral particle (neutron and photons) kinematics at extremely forward angles covering polar angles θ<5.5 mrad(η>6). Subsequently, this enables to study the three-dimensional structure of nuclei and hadrons in the downstream analysis.
A reconstruction framework is developed by incorporating Boosted Decision Tree(BDT)-based regressor and classifier. Single particle (lambda, neutron) interaction with beam energy in the range from 50 GeV to 200 GeV are simulated into the Silicon-Tungsten ECAL and steel scintillator HCAL sampling calorimeters. The BDT regressor reconstructs the photon and neutron shower energy whereas the classifier used for particle identification. We have explored various hand-engineered features to identify the feature with most expressing power
in PID. The classifier using shower-shape variables separates Λ from neutron events with a ROC-AUC of 0.986 ± 0.002 and 94.6% accuracy. It is observed that the discriminating power comes from the lateral shower shape and position, not from timing.
The obtained result will further serve as a benchmark to evaluate the performance gain of a Graph Neural Network(GNN)-based algorithm developed for the PID and shower energy reconstruction. In this work, we present the comparative study of shower reconstruction and particle identification using both BDT and GNN algorithms.