Applying Machine Learning to GlueX Data Analysis

GlueX is a high energy physics experiment with the goal of collecting data necessary for understanding confinement in quantum chromodynamics. Beginning in 2015, GlueX will collect huge amounts of data describing billions of particle collisions. In preparation for data collection, efforts are underway to develop a methodology for analyzing these large data sets. One of the primary challenges in GlueX data analysis is isolating events of interest from a proportionally large background. GlueX has recently begun approaching this selection problem using machine learning algorithms, specifically boosted decision trees. Preliminary studies indicate that these algorithms have the potential to offer vast improvements in both signal selection efficiency and purity over more traditional techniques.