Performance and optimization of support vector machines in high-energy physics classification problems

Abstract

In this paper we promote the use of Support Vector Machines (SVM) as a machine learning tool for searches in high-energy physics. As an example for a new- physics search we discuss the popular case of Supersymmetry at the Large Hadron Collider. We demonstrate that the SVM is a valuable tool and show that an automated discovery- significance based optimization of the SVM hyper-parameters is a highly efficient way to prepare an SVM for such applications. A new C++ LIBSVM interface called SVM-HINT is developed and available on Github.

20 pages, 6 figures

References in corpus (11)

• PYTHIA 6.4 Physics and Manual
• Implications of CTEQ global analysis for collider observables
• MCFM for the Tevatron and the LHC
• ROOT - A C++ Framework for Petabyte Data Storage, Statistical Analysis and Visualization
• Simplified Models for a First Characterization of New Physics at the LHC
• Squark and gluino production cross sections in pp collisions at $\sqrt{s}$ = 13, 14, 33 and 100 TeV
• Evaluation of three methods for calculating statistical significance when incorporating a systematic uncertainty into a test of the background-only hypothesis for a Poisson process
• ATLAS Run 1 searches for direct pair production of third-generation squarks at the Large Hadron Collider
• Non-Simplified SUSY: Stau-Coannihilation at LHC and ILC
• Support Vector Machine Classification on a Biased Training Set: Multi-Jet Background Rejection at Hadron Colliders
• Search for the standard model Higgs boson produced in association with a $W^{\pm }$ boson with 7.5 fb${}^{-1}$ integrated luminosity at CDF

Cited by in corpus (1)

• Support Vector Machines and generalisation in HEP