Predicting Z Boson Decay Modes: Evaluating the Performance of Machine Learning and Deep Learning Techniques in Particle Physics
Canda Deniz Sag, Onur Sahin · 2023
This study presents an application of machine learning and deep learning techniques to the problem of classifying the decay modes of the Z boson, a fundamental particle in the Standard Model of particle physics. We focus on two specific decay modes: Zee and Zmumu. Our dataset comprises simulated events with variables such as transverse momentum, pseudorapidity, and phi angle of the decay leptons. The data is preprocessed via a mapping method for categorical data and a MinMaxScaler for numerical data. We compare the performance of several classification models, including K-Nearest Neighbors (KNN), Weighted KNN, Random Forest Classifier, KDTree KNN, XGBoost, Multi Layer Perceptron, Support Vector Machine, Decision Tree Classifier, and a Deep Learning model based on a sequential neural network architecture. Our results indicate that the XGBoost provides the highest accuracy (81.42%) among the tested models. However, the Deep Learning model exhibits superior precision (84.14%), which is crucial in reducing false positives in the classification task. The study emphasizes the potential of machine learning and deep learning in high-energy physics and suggests directions for future research to further improve the classification performance.