?
Анализ аффективной поляризации в российском онлайн-пространстве в условиях политической мобилизации: подход машинного обучения
In the context of the current rise in global political tensions and widespread protests, the examination of affective polarization dynamics has become increasingly significant. This phenomenon, which previously dominated the American political scene and was studied through the lens of party affiliation, is now gaining attention in other countries as well. Affective polarization has been observed in Russia, a nation with a long history of political cleavages. Despite the significance of this problem, research on affective polarization presents challenges in developing unbiased and objective methods for analysis that are free from traditional survey biases. Additionally, it remains an open question whether affective polarization and political activism reinforce each other, which is the central hypothesis of my research. To evaluate the proposed methodology, a study based on machine learning techniques is conducted. Specifically, the following methods are employed: spectral clustering, thematic modeling using BERTopic, and Shannon entropy analysis. As a basis for the empirical analysis, primary data from the VKontakte social media platform during the summer protests in Moscow in 2019 is utilized. These data consist of user- generated messages containing instances of hate speech, which serve as a primary indication of affective polarization among participants. The findings support the hypothesis that there was a statistically significant increase in the level of Shannon entropy, as well as the emergence of polarized themes associated with protest mobilization. These results suggest that political mobilization may be linked to the propagation of affective polarization within society. Through the application of the developed methodology, an objective assessment of political processes can be achieved. Furthermore, this approach provides a means for monitoring and evaluating risks associated with escalating social tensions.