Analiza sentimenta u tvitovima na hrvatskome jeziku

master's thesis

• Title (english): Sentiment Analysis in Tweets in Croatian Language
• Author: Luka Krajcar
• Mentor: Jan Šnajder (mentor)
• Committee member: Jan Šnajder (predsjednik povjerenstva)
• Committee member: Marin Golub (član povjerenstva)
• Committee member: Nikola Bogunović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-08, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U posljednjih nekoliko godina primjetan je nagli porast korisnički generiranog sadržaja
unutar društvenih mreža. Jedna od takvih mreža osobito pogodna za strojnu
analizu sentimenta je Twitter, mikro-blogging servis unutar kojega korisnici odašilju
kratke poruke - tvitove. Cilj strojne analize sentimenta je automatski odrediti mišljenje, stav ili emociju izraženu u tekstu klasifikacijom u jednu od tri klase: pozitivnu,negativnu ili neutralnu. Zbog nemogućnosti ručnog označavanja velikog broja poruka koristi se model blago nadziranog strojnog učenja. U radu je ispitan utjecaj raznih metoda predobrade podataka te rad nekoliko klasifikacijskih modela poput naivnog Bayesovog klasifikatora, stroja potpornih vektora i logističke regresije. Prikazana je i obrazložena usporedba modela nadziranog i blago nadziranog strojnog učenja. Prikaz agregiranog sentimenta i statistike omogućen je kroz korisničku aplikaciju.

Abstract (english)

In recent years there has been a sharp rise in user-generated content within social
networks. One of these networks particularly suitable for machine analysis of sentiment
is Twitter, a micro-blogging service within which users broadcast short messages
- tweets. The goal of machine sentiment analysis is to automatically determine the
opinion, attitude or emotion expressed in text by classification into one of three categories:
positive, negative or neutral. Due to the inability of manual labeling of a
large number of messages a model of distant supervised machine learning. This paper
investigates the impact of different methods of preprocessing data and effectiveness of
several classification models such as naive Bayes classifier, support vector machines
and logistic regression. Comparison of supervised and distant supervised machine learning
is presented and explained in detail. Representation of aggregate sentiment and
statistics is provided via a user application.

• Keywords: Twitter; obrada prirodnog jezika; analiza sentimenta; strojno učenje; hrvatski jezik
• Keywords (english): Twitter; natural language processing; sentiment analysis; machine learning; croatian language
• Language: croatian
• URN:NBN: urn:nbn:hr:168:908184
• 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: 2018-02-06 09:52:45