Klasifikacija glazbenih djela po žanrovima

Bošnjak, Dominik

master's thesis

Klasifikacija glazbenih djela po žanrovima

Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2021. urn:nbn:hr:200:229480

Bošnjak, Dominik

Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Computer Engineering and Automation
Chair of Automation and Robotics

Institutional repository: Faculty of Electrical Engineering, Computer Science and Information Technology Osijek

Cite this document

APA 6th Edition

Bošnjak, D. (2021). Klasifikacija glazbenih djela po žanrovima (Master's thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:229480

MLA 8th Edition

Bošnjak, Dominik. "Klasifikacija glazbenih djela po žanrovima." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2021. https://urn.nsk.hr/urn:nbn:hr:200:229480

Chicago 17th Edition

Harvard

Bošnjak, D. (2021). 'Klasifikacija glazbenih djela po žanrovima', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 25 February 2025, https://urn.nsk.hr/urn:nbn:hr:200:229480

Vancouver

Bošnjak D. Klasifikacija glazbenih djela po žanrovima [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2021 [cited 2025 February 25] Available at: https://urn.nsk.hr/urn:nbn:hr:200:229480

IEEE

D. Bošnjak, "Klasifikacija glazbenih djela po žanrovima", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2021. Available at: https://urn.nsk.hr/urn:nbn:hr:200:229480

Cite this item: https://urn.nsk.hr/urn:nbn:hr:200:229480

Metadata

Title	Klasifikacija glazbenih djela po žanrovima
Title (english)	Music Genre Classification
Author	Dominik Bošnjak
Mentor	Emmanuel Karlo Nyarko (mentor)
Committee member	Emmanuel Karlo Nyarko (predsjednik povjerenstva)
Committee member	Ratko Grbić (član povjerenstva)
Committee member	Petra Pejić (član povjerenstva)
Granter	Josip Juraj Strossmayer University of Osijek Faculty of Electrical Engineering, Computer Science and Information Technology Osijek (Department of Computer Engineering and Automation) (Chair of Automation and Robotics) Osijek
Defense date and country	2021-10-05, Croatia
Scientific / art field, discipline and subdiscipline	TECHNICAL SCIENCES Computing Artificial Intelligence
Abstract	Kroz rad diskutiran je problem klasifikacije glazbenih djela po žanrovima. Postavljena je problematika te su sagledani mogući pristupi rješavanju problema. Dani su dostupni podatkovni skupovi za problematiku te je obrazložen proces izabiranja značajki koje bi mogle dostatno prikazati različite aspekte glazbe ovisno o žanru. Nadalje, uspoređeno je nekoliko mogućih metoda za klasifikaciju, konkretno metoda K najbližih susjeda, stroj sa potpornim vektorima te neuronske mreže. Za svaku od metoda pružen je kratki opis te proces treniranja. Nakon treniranja prikazani su dobiveni rezultati te je odrađena usporedba koja je pokazala uspješnost modela. Neuronske mreže pokazale su se najuspješnijima, za njima slijedi stroj sa potpornim vektorima te je „najlošije“ rezultate dala metoda K najbližih susjeda. Također, usporedbom je utvrđeno kako postoje određeni trendovi vezani za klasifikaciju žanrova pa je tako bez obzira na metodu, Klasična glazba uvijek bila najtočnije klasificirana, dok su najgori rezultati dobiveni za Rock žanr. U konačnici, pokazano je kako odabrane metode daju podjednake rezultate iz čega slijedi zaključak da za poboljšanje rezultata nisu potrebi novi, sofisticiraniji algoritmi i tehnologije, već da bi se fokus trebao prebaciti na kreiranje cjelovitijih podatkovnih skupova te na proces izvlačenja značajki koje bi vjernije prikazivale glazbeni žanr u pogledu numeričkih vrijednosti.
Abstract (english)	The problem of music genre classification is discussed in this paper. The scope of the problem is given along with possible solutions. Next, available datasets are listed and feature extraction process for sufficient music genre representation is explained. Furthermore, several classification methods are compared, specifically K Nearest Neighbors, Support Vector Machines and Neural Networks. For each method a short description along with training process is provided. Afterwards, obtained results are compared between methods. Neural Networks have proven to be most successful, followed by Support Vector Machines and finally K Nearest Neighbors with the lowest success rate. Likewise, comparison has shown that there are some trends within the dataset samples hence no matter which method was used, Classical music always had the highest, while Rock music showed the lowest accuracy amongst the genres. Ultimately it has been shown how the evaluated methods give similar results from which it can be concluded that for obtaining higher accuracies there is no need for more sophisticated algorithms or new technologies, but more effort needs to be invested in creating better datasets. Likewise, features for better musical genre representation could be considered.
Keywords
Keywords (english)
Language	croatian
URN:NBN	urn:nbn:hr:200:229480
Study programme	Title: Graduate University Study Programme in Computer Engineering Study programme type: university Study level: graduate Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resource	Text
File origin	Born digital
Access conditions	Open access
Terms of use
Created on	2021-11-24 12:54:11

Search form