Sustav za preporuku pjesama na temelju korisničkih ocjena

master's thesis

• Title (english): Song recommender system based on user ratings
• Author: Lucijan Peš
• Mentor: Marin Vuković (mentor)
• Committee member: Marin Vuković (predsjednik povjerenstva)
• Committee member: Dragan Jevtić (član povjerenstva)
• Committee member: Miljenko Mikuc (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U radu su opisane vrste sustava za preporuku sadržaja te su istražena i opisana postojeća rješenja sustava za preporuku pjesama na temelju njihovog sadržaja. Obrađene su metode latentne reprezentacije i redukcije dimenzionalnosti ulaznih podataka. Kao rezultat istraživanja predstavljen je i implementiran novi sustav preporuka na temelju sadržaja pjesama. Za određivanje pripadnosti pjesama pojedinim kategorijama koristio se algoritam gradient boosting koji se pokazao efektivnim pri učenju nad malim brojem primjera s velikim brojem značajki. Objektivnom i subjektivnom analizom rezultata pokazalo se kako modelirani sustav uspješno gradi liste pjesama koje uvelike odgovaraju korisničkim upitima te poboljšavaju korisničko iskustvo.

Abstract (english)

This paper goes through different types of content recommendation systems and explores existing content-based music recommender systems. Different latent representation methods and dimensionality reduction techniques of input data were presented. As a result of the research, a new content-based music recommender system was introduced and implemented. To determine the affiliation of songs to particular categories, a gradient boosting algorithm was used which proved to be effective in case of a small number of examples with a large number of features. Objective and subjective analysis of the results showed that the modeled system successfully builds song playlists that largely respond to user queries and improves overall user experience.

• Keywords: sustav za preporuku; latentna reprezentacija; redukcija dimenzionalnosti; gradient boosting; MVC
• Keywords (english): content-based music recommender system; latent representation; dimensionality reduction; gradient boosting; MVC
• Language: croatian
• URN:NBN: urn:nbn:hr:168:488415
• Study programme: Title: Computing; 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: Closed access
• Created on: 2020-01-07 21:16:09