Primjena strojnog učenja za tematsku analizu sentimenta

undergraduate thesis

• Title (english): Application of Machine Learning for Topic-Based Sentiment Analysis
• Author: Stipan Mikulić
• Mentor: Jan Šnajder (mentor)
• Committee member: Jan Šnajder (predsjednik povjerenstva)
• Committee member: Domagoj Alagić (član povjerenstva)
• Committee member: Mladen Karan (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2016-07-11, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Uslijed ogromnog povećanja korisnički generiranog sadržaja na društvenih mrežama,
detekcija tema i analiza sentimenta nameću se kao nezaobilazni alati za analizu javnog
mijenja. U radu su proučeni postupci za detekciju tema u tekstnih dokumenata te anal-
izu sentimenta prema detektiranim temama. Naglasak je stavljen na postupke temel-
jene na modelima strojnog učenja. Detekcijom tema nastoje se prepoznati teme o
kojima se priča na društvenim mrežama te analizom sentimenta se odred̄uje da li je
stav o tim temama pozitivan, negativan ili neutralan. Rad se sastoji od cijelog procesa
prikupljanja i preprocesiranja podataka. Nakon toga slijedi izgradnja modela detekcije
tema i analize sentimenta te evaluacija modela. Dodatno, za analizu sentimenta je bilo
potrebno označiti podatke za treniranje modela.

Abstract (english)

Due to the huge increase in user-generated content on social networks, topic detec-
tion and sentiment analysis imposed as indispensable tools for the analysis of public
opinion. The thesis examines the procedures for the topic detection in text documents
and an sentiment analysis towards the detected topics. Special focus is placed on a
procedures based on machine learning. Topic detection attempts to identify the topics
on which the story on social networks and sentiment analysis to determine whether the
position on these issues is positive, negative or neutral. The work consists of the en-
tire process of collecting and preprocessing data. This is followed by the construction
of topic detection and sentiment analysis models and evaluation of those models. In
addition, sentiment analysis required data annotation for training the model.

• Keywords: obrada prirodnog jezika; analiza sentimenta; strojno učenje; detekcija teme; latentna Dirichletova dodjela; stroj potpornih vektora; računalna lingvistika
• Keywords (english): natural language processing; sentiment analysis; machine learning; topic detection; latent Dirichlet allocation; support vector machines; Croatian language; computational linguistics
• Language: croatian
• URN:NBN: urn:nbn:hr:168:348565
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 10:43:46