izpis_h1_title_alt

Analiza tematske politične usmerjenosti slovenskih tvitov
ID KORELIČ, MARTIN (Author), ID Robnik Šikonja, Marko (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (3,06 MB)
MD5: 75126A190353CCEBB833F77AE1380BFA

Abstract
Družbena omrežja omogočajo prosto javno izražanje političnih mnenj uporabnikov, ki zagovarjajo različna stališča glede aktualnih političnih vprašanj. V diplomski nalogi smo analizirali politično usmerjenost oziroma pristranskost slovenskih uporabnikov na podlagi njihovih objav na družbenem omrežju Twitter. Pri tem smo uporabili metode za obdelavo naravnega jezika. Z uporabo algoritma BERTopic smo poiskali in iz podatkovne množice izluščili različne politično družbene teme in jih uporabili pri analizi sentimenta za klasifikacijo politične usmerjenosti (levo, desno, nevtralno). Opazimo precejšen delež negativnega sentimenta do vseh tem in strank. Količina levo in desno usmerjenih tvitov v političnih temah obeh polov je približno enaka. Zaznamo, da v tvitih po priljubljenosti najbolj izstopata dve stranki, vsaka iz nasprotnega političnega pola.

Language:Slovenian
Keywords:obdelava naravnega jezika, BERT, BERTopic, SloBERTa, politična mnenja, Twitter, analiza sentimenta
Work type:Bachelor thesis/paper
Typology:2.11 - Undergraduate Thesis
Organization:FRI - Faculty of Computer and Information Science
Year:2022
PID:20.500.12556/RUL-142496 This link opens in a new window
COBISS.SI-ID:129873411 This link opens in a new window
Publication date in RUL:11.11.2022
Views:918
Downloads:151
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Analysis of topical political stance of Slovene tweets
Abstract:
Social networks allow free public expression of users' political opinions, advocating various views on the current political agenda. In the thesis, we analyzed the political orientation of Slovene users' posts on the Twitter social network. We used the BERTopic algorithm to find and extract political topics from the data and applied sentiment analysis to classify political orientation (left, right and neutral). The results show a significant proportion of negative sentiment towards all topics and parties. The amount of left- and right-leaning tweets on general political topics is approximately equal. We notice that two parties from opposite political poles stand out in tweet popularity.

Keywords:natural language processing, BERT, BERTopic, SloBERTa, political opinions, Twitter, sentiment analysis

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back