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.
Computational Analysis of Polarization in Central Asian Digital Discourse: A Transformer-Based Approach to Discursive Strategies in Social Media Comments
DOI:
Keywords:
Digital discourse
Central Asia
Social media comments
Abstract
Downloads
Download data is not yet available.
References
Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding.
Conneau, A., et al. (2020). Cross-lingual Representation Learning at Scale.
Jurafsky, D., & Martin, J. H. (2021). Speech and Language Processing.
Sunstein, C. (2017). Divided Democracy in the Age of Social Media.
Tufekci, Z. (2015). Algorithmic Harms beyond Facebook and Google.
Абдуллаев, А. (2021). Медийное пространство Центральной Азии: язык и общество. Ташкент.
Каримов, Б. (2020). Социальные сети и общественное мнение в Узбекистане.
Нуртазин, М. (2019). Языковая политика и идентичность в Казахстане.
Rahimov, D. (2022). Digital Communication Trends in Uzbekistan.
Saidov, U. (2021). Интернет-дискурс в Узбекистане.