prikaz prve stranice dokumenta Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru
Access restricted to students and staff of home institution
undergraduate thesis
Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru
2021. urn:nbn:hr:179:862152
Mišić, Mijo
University of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
Department of Mathematics and Physics

Cite this document

Mišić, M. (2021). Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru (Undergraduate thesis). Retrieved from https://urn.nsk.hr/urn:nbn:hr:179:862152

Mišić, Mijo. "Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru." Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2021. https://urn.nsk.hr/urn:nbn:hr:179:862152

Mišić, Mijo. "Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru." Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2021. https://urn.nsk.hr/urn:nbn:hr:179:862152

Mišić, M. (2021). 'Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru', Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, accessed 22 March 2023, https://urn.nsk.hr/urn:nbn:hr:179:862152

Mišić M. Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru [Undergraduate thesis]. Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture; 2021 [cited 2023 March 22] Available at: https://urn.nsk.hr/urn:nbn:hr:179:862152

M. Mišić, "Korištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru", Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, Split, 2021. Available at: https://urn.nsk.hr/urn:nbn:hr:179:862152

Please login to the repository to save this object to your list.
Metadata
TitleKorištenje metoda strojnog učenja za analizu fizikalnih procesa na CMS detektoru
Title (english)Using machine learning methods to analyse physical processes at the CMS detector
AuthorMijo Mišić
MentorDamir Lelas (mentor)
GranterUniversity of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
(Department of Mathematics and Physics)
Split
Defense date and country2021-09-17, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Electrical Engineering
Abstract
S obimnom količinom podataka generiranih svaki dan u sudarima na Velikom sudaraču hadrona (LHC) u Ženevi, Švicarska, kao i zbog važnosti pronalaska određenih rijetkih signala, postoji neprekidna težnja za korištenjem sve naprednijih programskih rješenja za analizu prikupljenih podataka. Ovaj rad uspoređuje programske implementacije metoda strojnog učenja i pokušava dati odgovor na pitanje optimalnog izbora. Analizirane su 3 metode, od toga su 2 relativno nove (PyKeras i PyTorch) i ne postoje detaljna istraživanja izvedbe istih, te jedna starija metoda strojnog učenja (Gradient Boosted Decission Tree - BDTG). BDTG je predstavljao polazišnu točku, jer se trenutno dominantno koristi za ovaj tip analize LHC podataka. U obradi podataka dobivenih CMS detektorom, utvrđeno je da postoji mogućnost znatnog unaprjeđenjan u klasifikaciji događaja s PyKeras i PyTorch metodama u odnosu na izvedbu BDTG-a.
Abstract (english)
With the vast amount of data generated each day in collisions at the Large Hadron Collider (LHC) in Geneva, Switzerland, as well as the importance of finding certain rare signals, there is a continuing tendency to use increasingly advanced software solutions to analyze collected data. This paper compares program implementations of machine learning methods and attempts to answer the question of optimal choice. 3 methods were analyzed, of which 2 are relatively new (PyKeras and PyTorch) and there are no detailed studies of their performance, as well as one older method of machine learning (Gradient Boosted Decission Tree - BDTG). BDTG was the starting point, as it is currently predominantly used for some of the LHC data analyses. In processing the data obtained by the CMS detector, it was found that there is a possibility of significant improvement in the classification of events with PyKeras and PyTorch methods in comparison to the performance of BDTG.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:179:862152
Study programmeTitle: Electrical Engineering and Information Technology
Study programme type: university
Study level: undergraduate
Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to students and staff of home institution
Terms of useLicence In copyright icon
Created on2021-12-08 11:25:48