UNIVERSITY OF BUCHAREST
FACULTY OF PHYSICS

Guest
2025-08-21 0:30
 HOME     CONFERENCES     SEARCH            LOGIN     NEW USER     IMAGES   


Conference: Bucharest University Faculty of Physics 2025 Meeting


Section: Theoretical and Computational Physics, High-Energy Physics, Applied Mathematics


Title:
Cut-Based methods vs. Machine Learning data analysis in High Energy Physics


Authors:
Alexandru MOROSANU (1), Victoria-Adriana VARGA (1)


Affiliation:
1)Faculty of Physics, University of Bucharest, Atomiștilor 405, RO-077125, Măgurele, Romania


E-mail
victoria-adriana.varga@s.unibuc.ro alexandru.morosanu@s.unibuc.ro


Keywords:
Cut-based, Machine Learning, comparison, signal to background


Abstract:
We discuss and compare two of the most used data analysis methods in High Energy Physics (HEP): cut-based analysis and machine learning (ML) algorithms. We analyze signal-to-background selection techniques in very large and varied data sets. Cut-based methods are recommended for first-hand analysis of smaller samples. The event selection is done by consecutively evaluating the data sets, based on some thresholds imposed on various observables (energy, pseudorapidity, transverse momenta etc.). ML algorithms, despite their complexity, are better at handling larger samples. They search for correlations and patterns between events in order to build a multivariate discriminator that can differentiate signal from background processes. From scratch, it takes longer to develop such an algorithm in contrast to a cut-based one. However, if done correctly, it compensates with much better performance. Concluding this comparison, we will put forward a general idea on the advantages, disadvantages, efficiency, and potential applications of the methods discussed, providing a comprehensive overview.


References:

[1] TMVA Users Guide [https://root.cern.ch/download/doc/tmva/TMVAUsersGuide.pdf]

[2] Cesare Bini, Data analysis in Particle Physics (2013)

[3] Sezen Sekmen Kyungpook National University / CMS, Experimental Methods and Physics at the LHC -II (2020)

[4] O. Behnke, K. Kröninger, G. Schotts - Data Analysis in High Energy Physics- A practical guide to statistical methods (2013)



Acknowledgement:
We would like to thank our coordinators from the ATLAS department of IFIN-HH, for the proposal of doing this study, as an introduction in the field of HEP.