Automated collection of social network data to develop a factor model of network self-presentation

Abstract

On the basis of the conducted empirical research of information images of users, the leading components of network self-presentation were revealed: statistical, socio-demographic component, visual component and value-semantic component. The authors analyzed the hidden factors responsible for the formation of network self-presentation through the information image, studying the prognostic possibilities of social profile data analysis. Significant differences in the content of the information image and socio-psychological characteristics of users with different styles of network self-presentation were identified. Algorithms for collecting and processing open information from social network profiles, followed by factor analysis, as well as machine learning methods to determine the topics of communities and interesting pages to which users subscribe, are presented. The paper deals with ethical and legal issues of using data collection from user groups to create predictive models based on them, without notifying the users themselves. Also discusses theoretical issues of interdisciplinary design model information image, which becomes possible through multilateral analysis of the symbolic content of social network profile: linguistic and psychological evaluation of the semantic content of the content of socio-psychological overview of communication practices that are implemented in the network and parsing features of the network interface that specifies the structure of this profile.