Application of Gene Expression Programming in Improving the Event Selection of the Semi-leptonic Top Quark Pair Process

Abstract

Searches for Beyond the Standard Model physics require probing the Standard Model with increased precision. One way this can be achieved is by improving the accuracy of the event selection classifiers. Recently, Gene Expression Programming (GEP) has been shown to provide complex yet easy to interpret classifiers in various fields. Previous attempts to apply GEP to high-energy physics (HEP), though limited by computational power available, achieved classifier accuracy of up to 95\%. In this paper, we demonstrate that a selection algorithm optimized by GEP and applied to the top-quark pair production process' semi-leptonic decay channel enables the increase of data purity for already highly pure data. Moreover, we explain how adding penalty cuts to the purity fitness function allows adjusting the optimized classifier to the needs of a specific measurement in terms of the size of the selected event sample and data purity.

References in corpus (7)

• Precision Higgs Physics at CEPC
• Variational Autoencoders for New Physics Mining at the Large Hadron Collider
• Learning to Classify from Impure Samples with High-Dimensional Data
• Demonstration of background rejection using deep convolutional neural networks in the NEXT experiment
• Measurement of the mass difference between top quark and antiquark in pp collisions at sqrt(s) = 8 TeV
• Support Vector Machines in Analysis of Top Quark Production
• Search for ${\Lambda }_c^{+}\to p{K}^{+}{\pi }^{-}$ and $D_s^{+}\to K^{+}{K}^{+}{\pi }^{-}$ Using Genetic Programming Event Selection