Topic Wise Influence Maximisation based on fuzzy modelling, Sentiments, Engagement, Activity and Connectivity Indexes

People's interactions, communications, and engagement have all changed as a result of the rise of social media. These networks are vital for expanding scope and impact. People in these networks influence each other. Social influencers spread the knowledge in the network. Identification of such influencers is a challenging task. Generally, in past studies varied qualitative metrics like centrality, connectivity etc has been popularly used for identifying the influencers. In a network it has been noticed that every person interacts in the network in context to its own interest areas. He influences specific to his interest domains. Based on this belief, the aim of this paper is to detect topic-wise influencers in social media so that we can target person or influencer appropriately for influence maximisation. This paper focuses on measuring the strength of topic-level social influence using sentiments of the text used for interaction by social network user and later fuzzy modeling has been applied. Fuzzy modeling help in finding the person probability index of influence (positive or negative) in context to a different topics he is contributing in social media. In addition, three user features- engagement Index, activity index and connectivity Index has been utilized to compute the user overall influencer score. For experimentation, the tweets from the Twitter have been used to evaluate the proposed method.