Identification of dielectron pairs with machine learning method

Öz

Dielectrons, electron (e-) positron (e+) pairs, are electromagnetic signals produced in various processes of high-energy particle collision experiments to understand the formation of the universe. Since these particle pairs do not interact strongly, they are not affected by the features of their environment. Therefore, they provide significant information about various production mechanisms. High purity pair signals are needed to measure dielectrons. Complex analysis techniques are required to distinguish these signals from much larger background sources (e+e+, e-e-). With conventional particle analysis methods, dielectron pairs are produced with high systematic uncertainties. In recent years, artificial intelligence (AI) applications in various fields have gained importance to increase the speed, accuracy and efficiency of human labors. In this study, artificial intelligence-based machine learning approach was used in dielectron analysis. In the study, the random forest (RO) classification method was applied to obtain dielectrons in the Large Hadron Collider 2010 data. In the study, the RO method applied dielectron analysis showed 98.2% success with 90.9% efficiency and 92.0% precision.

Anahtar Kelimeler

• Dielectron
• random forest classifier
• machine learning

Makine öğrenmesi yöntemi ile dielektron çiftlerinin tanımlanması

Öz

Dielektronlar olarak adlandırılan elektron (e-) pozitron (e+) çiftleri, evrenin oluşumunu anlamak için yapılan yüksek enerjili parçacık çarpışma deneylerinin çeşitli süreçlerinde üretilen elektromanyetik sinyallerdir. Bu parçacık çiftleri, güçlü kuvvet etkileşimi yapmamaları sebebiyle bulundukları ortamın özelliklerinden etkilenmezler ve böylece çeşitli üretim mekanizmaları ile ilgili önemli bilgi sağlarlar. Dielektronları ölçmek için yüksek saflıkta çift sinyalleri gereklidir. Bu sinyalleri, kendisinden çok daha büyük olan arka plan (fon) kaynaklarından (e+e+, e-e-) ayırt etmek için karmaşık analiz teknikleri gereklidir. Geleneksel parçacık analiz yöntemleri ile dielektron çiftleri yüksek sistematik belirsizlikler ile üretilir. Son yıllarda çeşitli alanlardaki yapay zeka (AI) uygulamaları, insan çabalarının hızını, doğruluğunu ve verimliliğini artırmak için önem kazanmaktadır. Bu çalışmada dielektron analizinde yapay zeka tabanlı makine öğrenmesi yaklaşımı kullanılmıştır. Çalışmada rastgele orman (RO) sınıflandırma yöntemi Büyük Hadron Çarpıştırıcısı 2010 yılı verisinde bulunan dielektronların elde edilmesine uygulanmıştır. Yapılan çalışmada %90.9 duyarlılık ve %92.0 kesinlik ile RO metodu uygulanmış dielektron analizleri %98.2 başarı göstermiştir.

Anahtar Kelimeler

• Dielektron
• rastgele orman sınıflandırma yöntemi
• makine öğrenimi.

Kaynakça

1. Wong, S. S. M., Introductory nuclear physics, 29, Wiley-VCH Press, Weinheim, (2004).
2. Griffiths, D. J., Introduction to elementary particles, 74, Wiley-VCH Press, Weinheim, (2008).
3. CMS Collab., Performance of electron reconstruction and selection with the CMS detector in proton-proton collisions at √s=8 TeV, JINST, 10, P06005, (2015).
4. CMS Collab, Measurement of the Inclusive W and Z Production Cross Sections in pp Collisions at √s = 7 TeV, JHEP, 10, 132, (2011).
5. Drell, S. D. ve Yan, T. M., Massive lepton-pair production in hadron-hadron collisions at high energies, Physical Review Letters, 25, 316, (1970).
6. Biernat, J., Measuring di-electron Dalitz decays of baryon resonances with HADES and PANDA, Doktora Tezi, Jagiellonian Üniversitesi, Nükleer Fizik Enstitüsü, Krakow, (2017).
7. Nourbakhsh, S., Studio degli eventi J/ in due elettroni con i primi dati di CMS, Doktora Tezi, Roma La Sapienza Üniversitesi, Matematik, Fizik ve Doğa Bilimleri Fakültesi, Roma, (2010).
8. ALICE Collab., Dielectron and heavy-quark production in inelastic and high-multiplicity proton–proton collisions at √s=13TeV, Physics Letters B, 788, 505, (2019).

9. STAR Collab., Measurements of Dielectron Production in Au+Au Collisions at √sNN =200 GeV from the STAR Experiment, Phys. Rev. C, 92, 024912, (2015).
10. ALICE Collab., Measurement of dielectron production in central Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Rev. C, 99, 024002, (2019).
11. STAR Collab., J/ψ production at high transverse momenta in p+p and Cu+Cu collisions at √sNN=200 GeV, Phys. Rev. C, 80, 041902, (2009).
12. ALICE Collab., J/ψ production as a function of charged-particle pseudorapidity density in p–Pb collisions at √sNN=5.02TeV, Physics Letters B, 776, 91, (2018).
13. ALICE Collab., Dielectron production in proton-proton collisions at √s=7 TeV, JHEP, 64, 1809, (2018).
14. Schwartz M. D., Modern Machine Learning and Particle Physics. arXiv:2103.12226, 2021.
15. Chen T., He T., Higgs Boson Discovery with Boosted Trees, Proceedings of the 2014 International Conference on High-Energy Physics and Machine Learning, 69-80, Montreal, (2014).
16. CMS Collab., Machine Learning Techniques in the CMS Search for Higgs Decays to Dimuons, Proceedings of 23rd International Conference on Computing in High Energy and Nuclear Physics, 06002, Sofia, (2019).
17. Arpaia P., Azzopardi G., Blanc F., Bregliozzi G., Buffat X., Coyle L., et al., Machine learning for beam dynamics studies at the CERN Large Hadron Collider, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 985, 164652, (2021).
18. Breiman, L., Random Forests, Machine Learning, 45, 5–32, (2001).
19. Trzcinski, T., Graczykowski, L. K. ve Glinka, M., Using Random Forest Classifier for particle identification in the ALICE Experiment, Proceedings of Information Technology, Systems Research and Computational Physics, 3-17, Krakow, (2019).
20. Trzcinski T. and Deja K., Assigning Quality Labels in the High-energy Physics Experiment ALICE Using Machine Learning Algorithms, Proceedings of NICA days, 647-655, Warsaw, (2017).
21. Müller A. C. ve Guido, S., Introduction to Machine Learning with Python, 28-30, O'Reilly Media Inc., Sebastopol CA, (2016).
22. Azhari, M., Alaoui, A., Achraoui, Z., Ettaki, B. ve Zerouaoui, J., Adaptation of the Random Forest Method, Proceedings of the 4th International Conference on Smart City Applications - SCA ’19, 1141–1146, Warsaw, (2019).
23. Azhari, M., Alaoui, A., Abarda, A., Ettaki, B. ve Zerouaoui , J., Big Data and Networks Technologies, 183-189, Springer 81, (2020).
24. Azhari, M., Alaoui, A., Abarda A., Ettaki, B. ve Zerouaoui, J., A Comparison of Random Forest Methods for Solving the Problem of Pulsar Search, Proceedings of the Fourth International Conference on Smart City Applications, 1-6, Cham, (2020).
25. McCauley, T., Events with two electrons from 2010, CERN Open Data Portal, (2014). https://opendata.cern.ch/record/304, (30.08.2021).
26. Pedregosa, F., Varoquaux G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M. ve Duchesnay, E., Scikit-learn: Machine learning in Python, Journal of Machine Learning Research, 12, 2825-2830, (2011).
27. NA61/SHINE Collab., Two-particle correlations in azimuthal angle and pseudorapidity in inelastic p+p interactions at the CERN Super Proton Synchrotron, Eur. Phys. J. C., 77, 59, (2017).
28. Bradley, A. P., The use of the area under the ROC curve in the evaluation of machine learning algorithms, Pattern Recognition, 30, 1145-1159, (1997).