Linguistic profiles of hidden communities: A morphosyntactic aspect

Abstract

The aim of the research is to identify quantitative regularities in the functioning of morphosyntactic parameters in the texts by users of hidden online communities. Through statistical methods, the paper attempts to confirm the “cohesion” of the main morphosyntactic features, the information about which was obtained using the Profiling-UD linguistic processor. The scientific novelty of the research lies in the following: based on a corpus of Russian-language social media texts, an experiment is conducted on the correlation analysis of morphosyntactic characteristics, which could become part of the future linguistic profile of hidden communities. Such profiles could be used in modern social media to enhance the functionality of recommendation systems. As a result, the research found that significant positive correlations with moderate statistical significance were identified for over 55% of hidden communities. By applying the proposed methodology, the linguistic profile of hidden communities can be further expanded with syntactic and lexical parameters, allowing for cluster analysis of communities and identification of the homogeneity/heterogeneity of the use of the characteristics across different linguistic levels in user posts from hidden communities.