Energy Distribution of Particle Flow in PP Collider: Evidenced by BP and MLP Neural Networks Prediction

Particle flow analysis is pivotal in high-energy physics experiments, especially for the examination of particle collisions. By combining calorimeter data from neutral and charged particles, this paper explores energy distributions within proton-proton (PP) collisions, employing machine learning techniques. Specifically, data from the Inner Tracker,Electromagnetic Calorimeter, and Hadronic Calorimeter are utilized to generate predictive graphs, which are then compared to reconstructed truth-level data. This approach establishes a benchmark for assessing the accuracy of particle flow analysis techniques. Furthermore, we demonstrate the application of backpropagation and multilayer perceptron (MLP) neural networks to enhance particle identification,incorporating magnetic field-induced deflection for improved prediction accuracy.