A Comparative Study of Neural Networks and Logistic Regression for High Energy Physics

Spring 2017

Elizabeth City State University

Keywords: Neural Networks, High Energy Physics, Logistic Regression

Mentor: Jerome Mitchell

This research was centered on evaluating the effectiveness of backpropagation and linear regression, two machine learning algorithms that were utilized in classifying the supersymmetry dataset retrieved from Large Hydron Collider. Machine learning is a valuable field in computer science that centers on algorithms that adapt to better perform a given task through large amounts of data. The supersymmetry dataset was a collection of data that was observed from particle collisions in addressing the possibility of the supersymmetry theory of particle physics being true. The supersymmetry theory suggests that the particles that have yet to be identified to complete the Standard Model are counterparts that mirror the current particles that have been discovered. In performing this research, the goal was to see which machine learning algorithm would stand as a better tool for classifying the data in the supersymmetry dataset.

URL: http://nia.ecsu.edu/ur/1617/teams/neural/