POSITIONING TWITTER USERS ON THE POLITICAL SPECTRUM BASED ON THE CONTENTS OF THEIR TWEETS
DOI:
https://doi.org/10.24867/20BE06KnezevicKeywords:
Twitter, political orientation, SVM, BERTAbstract
This paper presents an approach to determine the political orientation of Twitter users based on the contents of their public available tweets. The approach is based on machine learning with the use of the Support Vector Machines classifier, BERT language model, and the Selenium library.
References
[1] Ideology Detection for Twitter Users via Link Analysis: Yupeng Gu, Ting Chen, Yizhou Sun and Bingyu Wang (http://web.cs.ucla.edu/~yzsun/papers/2017_SBP_Ideology)
[2] Selenium WebDriver (https://www.selenium.dev/documentation/webdriver)
[3] BERTić - The Transformer Language Model for Bosnian, Croatian, Montenegrin and Serbian, Nikola Ljubešić, Davor Lauc (https://aclanthology.org/2021.bsnlp-1.5.pdf)
[4] SrbAI - Python biblioteka za procesiranje srpskog jezika (https://github.com/Serbian-AI-Society/SrbAI)
[5] Support Vector Machines – Scikit (https://scikit-learn.org/stable/modules/svm.html)
[6] MatPlotLib (https://matplotlib.org)
[2] Selenium WebDriver (https://www.selenium.dev/documentation/webdriver)
[3] BERTić - The Transformer Language Model for Bosnian, Croatian, Montenegrin and Serbian, Nikola Ljubešić, Davor Lauc (https://aclanthology.org/2021.bsnlp-1.5.pdf)
[4] SrbAI - Python biblioteka za procesiranje srpskog jezika (https://github.com/Serbian-AI-Society/SrbAI)
[5] Support Vector Machines – Scikit (https://scikit-learn.org/stable/modules/svm.html)
[6] MatPlotLib (https://matplotlib.org)
Downloads
Published
2022-11-02
Issue
Section
Electrotechnical and Computer Engineering
