An Agent-Network Approach to Forecasting Political Activity of Russian Social Media Users

This article systematizes the experience of applying an agent-network approach to forecasting social media political activity of digital platform users. Based on a review of foreign and Russian scientific experience, the author identifies key opportunities and limitations of using agent-based modeling (ABM) in the Russian information environment. The paper establishes that Western methodologies (agent-based modeling, network models of information diffusion, forecasting based on deep neural networks) demonstrate high effectiveness in predicting political activity in open information spaces. However, their adaptation to the Russian context requires consideration of the specifics of state regulation of the information field, the architecture of social media platforms, and the structure of Russian public opinion. The author determines the optimal methods for model parameterization, and proposes a comprehensive approach to constructing political activity forecasting systems, integrating the methodological tools of Western computational science with Russian experience in analyzing information flows and network structures.