Computational Analysis of Polarization in Central Asian Digital Discourse: A Transformer-Based Approach to Discursive Strategies in Social Media Comments

Abstract

Paper examines the nature of polarization in Central Asian social media discourse using transformer-based computational models. It combines quantitative methods, such as sentiment and toxicity analysis, with qualitative discourse analysis to identify key communicative strategies. The findings reveal a high level of negative and conflict-oriented language, often shaped by multilingual and socio-political contexts. The study highlights both the effectiveness and limitations of NLP tools in capturing culturally specific meanings in digital communication.

Topics

Identifiers

DOI

10.37547/ajsshr/volume06issue04-12

Citations and references