Socio-political News Spread in Russian Language Online Social Network

Project leader: Sergei Koltsov

Project participants: Sophia Dokuka, Sergei Pashakhin, Maksim Koltsov, Vera Ignatenko

2018—2020

The project is supported with RFBR grant N 18-011-00997.

The goal of this research is to identify mechanisms that drive dissemination of professionally produced public affairs news across a Russian-language online social network (represented by VKontakte). The research investigates which features of messages, users and network topology influence this process, and which of them contribute to emergence of information cascades. The main anticipated result of this project is a model describing parameters of such cascades and the knowledge about the role of various factors in their generation. The research exploits a large dataset of text and user data from VKontakte and applies mathematical diffusion models along with automatic text analysis. The project is unique for the Russian language online social networks that have never been an object of a study into information cascading. Furthermore, to the best of our knowledge, this research is the first to describe dissemination of full-text professional news (not tweets) and the first to ask which factors influence users’ decisions to alter news content by adding their utterances. The practical importance of such study is obvious from the growing role of social networks in news dissemination that increasingly influence behavior of large populations. Understanding mechanisms of news diffusion through online networks may help manage news flows more efficiently both for protecting users from malicious content, such as fraudulent information, and for delivering socially useful news to target audiences.

The project aims to study 12 TV channels present in VKontakte social network. During the first year, it is planned to collect data and analyze agenda differences and similarities in TV channels content using topic modeling. Additionally, this phase includes sentiment analysis of posts and comments as well as cluster analysis of TV channels' auditories.

Publication:

• Koltsov, S., Pashakhin, S., & Dokuka, S. (2018). A Full-Cycle Methodology for News Topic Modeling and User Feedback Research. In S. Staab, O. Koltsova, & D. I. Ignatov (Eds.), Social Informatics (pp. 308–321). Springer International Publishing.  https://doi.org/10.1007/978-3-030-01129-1_19  Download
• Koltsova, O. Y., Dokuka, S., Koltsov, S., & Koltcov, M. (2018). (in print) Echo chambers vs opinion crossroads in news consumption on social media. In W. M. van der Aalst, D. I. Ignatov, M. Y. Khachay, S. Kuznetsov, V. Lempitsky, I. A. Lomazova, … S. Wasserman (Eds.), Analysis of Images, Social Networks and Texts. 7th International Conference, 2018, Lecture Notes in Computer Science, Revised Selected Papers (Vol. 11179, pp. 7–14). Saint Petersburg: Springer. Download preprint version
•  S. Koltcov, V. Ignatenko, S. Pashakhin. An Entropic approach to the problem of determining the optimal number of clusters in hierarchical cluster analysis. Technical Physics Letters. (under review)