prikaz prve stranice dokumenta Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora
No public access
master's thesis
Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2019. urn:nbn:hr:168:611728
Volarić Horvat, Leonard
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Volarić Horvat, L. (2019). Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora (Master's thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:611728

Volarić Horvat, Leonard. "Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2019. https://urn.nsk.hr/urn:nbn:hr:168:611728

Volarić Horvat, Leonard. "Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2019. https://urn.nsk.hr/urn:nbn:hr:168:611728

Volarić Horvat, L. (2019). 'Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora', Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 01 June 2024, https://urn.nsk.hr/urn:nbn:hr:168:611728

Volarić Horvat L. Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora [Master's thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2019 [cited 2024 June 01] Available at: https://urn.nsk.hr/urn:nbn:hr:168:611728

L. Volarić Horvat, "Računalna stilometrijska analiza transkripata rasprava Hrvatskog sabora", Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2019. Available at: https://urn.nsk.hr/urn:nbn:hr:168:611728

Metadata
TitleRačunalna stilometrijska analiza transkripata rasprava Hrvatskog sabora
Title (english)Computational Stylometry Analysis of Croatian Parliamentary Discussions
AuthorLeonard Volarić Horvat
MentorJan Šnajder (mentor)
Committee memberJan Šnajder (predsjednik povjerenstva)
Committee memberSlobodan Ribarić (član povjerenstva)
Committee memberAlan Jović (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2019-07-08, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Jedna od najvažnijih karakteristika svake politički angažirane osobe je retorika koju koristi. Retorika se sastoji od sadržaja i stila: načina na koji se sadržaj prenosi javnosti. Ove dvije komponente retorike često bivaju istraživane u lingvistici, psiholingvistici i političkim znanostima, a zbog napretka računalne obrade prirodnog jezika, alati razvijeni u računalnoj lingvistici se sve češće koriste u ovakvim analizama. Većina prethodnih istraživanja stilskih značajki političkog teksta odnosi se na tekstove na engleskom jeziku. Ovaj rad proučava rasprave Hrvatskog sabora, i to stilometriju prisutnu u transkriptima tih raspravi. Izvlačenje stilskih značajki iz teksta pokazalo se zahtjevnim, budući da je riječ o govornom jeziku, čije stilometrijske karakteristike nisu toliko istražene koliko jesu u pisanom jeziku, a pogotovo za hrvatski jezik. Nakon izvlačenja stilskih značajki i skupljanja dodatnih podataka iz drugih izvora, eksplorativna je analiza ponudila uvid u statističke karakteristike odabranih značajki. Također, koristeći algoritam t-SNE, podatci su iscrtani u dvije dimenzije, što je otkrilo nekoliko grupa među različitim skupovima govornika. Nakon toga, obavljena je regresijska analiza podataka, s ciljem stvaranja modela za procjenu političke orijentacije pojedinog saborskog zastupnika, na temelju dostupnih stilskih značajki.
Abstract (english)
Any politician’s main tool is their rhetoric, which consists of content and style – the way that content is transmitted to the public. The two components have been subjected to previous research, primarily in the domain of Linguistics, Psycholinguistics, and Political Science. Due to recent advances in Natural Language Processing, analysis of political text relies more and more on tools developed within Computational Linguistics. Most previous work in stylometry of political text was done in English. This thesis takes a closer look at the discussions in the Croatian Parliament, by examining the stylometry in the transcripts of said discussions. Extracting stylometric features from the transcripts proved challenging, because analyzing transcripts spoken language is not a task often encountered in the research community, and even less so in Croatian. After extracting stylometric features, and combining the data with several external data sources, an exploratory analysis was done. Along with basic descriptive statistics, the data was visualized using t-SNE, a dimensionality reduction technique, which revealed several clusters within different subsets of the data. After exploring the data, a regression analysis was performed, with the aim of developing a model for predicting the political orientation of a speaker, based on their stylometric features.
Keywords
Keywords (english)
Languageenglish
URN:NBNurn:nbn:hr:168:611728
Study programmeTitle: Computing
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2020-01-07 22:56:43