Klasifikacija pjesama u žanrove na temelju stihova

master's thesis

• Title (english): Genre Classification of Songs Based on Lyrics
• Author: Antonijo Marijić
• Mentor: Marina Bagić Babac (mentor)
• Committee member: Marina Bagić Babac (predsjednik povjerenstva)
• Committee member: Mirjana Domazet-Lošo (član povjerenstva)
• Committee member: Marko Horvat (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-06, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Klasifikacija pjesama u žanrove na temelju stihova zahtjevan je zadatak čak i za ljude, no suvremena obrada prirodnog jezika nedavno je ponudila napredna rješenja ovog zadatka. Ovaj diplomski rad obuhvaća predobradu skupa podataka o stihovima, te implementira i daje usporedbu različitih modela strojnog učenja uz utjecaj različitih načina vektorizacije riječi na rezultate (GloVe, Word2Vec, kreiranje rječnika i korištenje frekvencija za kreiranje vektora), u rasponu od logističke regresije, stroja potpornih vektora i slučajne šume kao referentnih modela, do modela dubokog učenja, koji obuhvaćaju neuronske mreže i transformere - RNN, LSTM, BiLSTM, BERT.

Abstract (english)

Genre classification of songs based on lyrics is a challenging task even for humans, however state-of-the-art natural language processing has recently offered advanced solutions to this task. In this thesis, preprocessing of the lyrics dataset is performed, and implementation and comparison of various machine learning models are provided along with the impact of different word embeddings on results (GloVe, Word2Vec, creating a dictionary and using frequency for the creation of vectors), ranging from logistic regression, support vector machine, random forest as benchmark models, to deep learning models, that are neural networks and transformers - RNN, LSTM, BiLSTM, BERT.

• Keywords: klasifikacija u žanrove; transformer; LSTM; vektorizacija riječi
• Keywords (english): genre classification; transformer; LSTM; word embedding
• Language: english
• URN:NBN: urn:nbn:hr:168:891036
• 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: 2023-01-14 09:26:08