New horizons of digital politics: linguistic analysis and neural networks in the study of socio-political processes

Abstract

The article presents an overview of a universal methodology for analyzing socio-political processes, based on the integration of the network approach (network analysis), mathematical and network modeling methods, and comprehensive linguistic analysis, which together form the basis of network linguistics - a fundamentally new research direction being developed by the Interdisciplinary Laboratory of Applied Linguistics and Network Studies of Public Policy at Kuban State University. The proposed methodology is modular in nature, incorporating methods that can be applied individually or collectively to examine various aspects of socio-political interaction. Its versatility allows for adaptation depending on the research objectives and the available tools. The study explores the synergy between traditional and modern analytical methods, including the semiotic analysis of hashtags and emojis, 3-D modeling of discursive fields, analysis of discursive dynamics, and the concept-sphere core. Special attention is given to graph neural networks as tools for the automated analysis of big data and for predicting the evolution of narratives based on an investigation of the predicative network within the discursive field. The presented analytical cases confirm the objectivity of the proposed approach and illustrate the in-depth analysis of digital discourses and their dynamics in political communication. Overall, the article elucidates the prospects of network linguistics in researching socio-political processes, especially considering the increasing volume of digital data and the necessity for precise predictive analytics tools.