Discovering Topic-Oriented Focal Sets in Cyber-Argumentation Using Link Analysis, Topic Modeling and Social Roles

Abstract

Discussions have implicit topics and are often exchanged by users with specific profiles. Users may work cooperatively or collectively to support, attack, or deliver agendas due to similar interests. Much research has been conducted to detect groups with similar interests or specific characteristics using link analysis techniques, text analysis techniques, or a combination of different methods on social media and blogs. However, most of the research that has been published has focused on improving the community, or hidden community, detection algorithms concerning the research challenges. In this research, a framework is proposed to discover groups, or focal sets, with similar topic interests, using the focal structure analysis algorithm and topic modeling in cyber argumentation; then, the social roles of the focal set members are investigated. By combining these techniques, we discover and examine groups and individuals behind specific topics in the discussion. In cyber-argumentation discussions, we can analyze the group’s and individuals’ structures and profiles. This work can identify groups and individuals behind specific topics and use their characteristics to blend communities and individuals of polarized opinions in an online discussion. This allows for balancing the groups and individuals in a discussion and draws out the crowd's wisdom in cyber-argumentation platforms.