Entropy based Weighted Features for Detecting the Influential Users on Twitter

Abstract

The research in influence social of social networks has attracted big interest in last years of its importance in marketing, information diffusion, and recommendation. Identifying which users have the power to influence the choices of other users is a significant research topic because it provides an opportunity for companies to identify influential users. However, most current methods to influencers find out by relying on measures of centrality computed on networks that use different kinds of interrelationships to link users. Popular users which have a large number of followers are not necessarily influential in terms of spawning mentions or retweets. A small proportion of all users are called the influencer set. Where these influencers have a big active audience that doubles the content and consumes published. For that reason, an influencer's post whether it is links or text is spread in the social network and draws the attention of a big number of individuals although they may not be direct followers to the influencer. The greater the spread of the post, the user impact is rise. In this research, the proposed method displays the ability to find influencer users by using eleven features such as mention, retweet actions, new tweets, the rate of followers to friends, and the count of public lists etc. which are completely effective indicators for influential users, by employing the Entropy method technique and threshold value Analysis for identifying users influential on Twitter.