Методологические основы анализа ценностных ориентаций в российском сегменте социальных сетей

Медиалингвистика. 2026. Т. 2. № 13.

Zverev S., Соколова Е. Н., Grigoreva M., Smirnov A.

The article presents a comprehensive methodology for analysing value orientations within the Russian segment of social networks, validated on a corpus exceeding two million posts. All research stages are described, from data collection using the Brand Analytics monitoring system and lexical preprocessing of texts (normalisation and cleaning) to topic modelling with Latent Dirichlet Allocation (LDA), clustering of value‐related topics (word–embedding representations combined with the k–means algorithm), and subsequent manual validation of the resulting clusters. The analysis yields a “map” of 25 value clusters, among which patriotism, love, family, mutual aid, and support are dominant. The findings reveal the prevalence of social values (patriotism, family, support) alongside a considerable share of self–actualisation values (success, achievement). The significance of the results for the sociology of values is discussed, along with practical applications: monitoring the dynamics of societal value orientations, informing information policy, advancing the national goal of fostering patriotic and socially responsible individuals, and designing educational programmes and communication strategies.

Keywords: ценностные ориентации, социальные сети, кластеризация, value orientations, clusterization, latent Dirichlet allocation, Latent Dirichlet Allocation (LDA), тематическое моделирование, social nets, topic modelling

Publication based on the results of: Modern Technologies of Socio-Political Communication (2025)