IDENTIFYING INFLUENTIAL SPREADERS IN COMPLEX NETWORKS

Abstract

One thing that we can't disregard in this day and age is incorporation of social media in everyday life. Because of expansion in reach of information to people in developing nations, consistently huge number of individuals access social network, a lot of them do that for the first time. Impact of social media is so much that it can affect one's point of view towards an issue. Since each social media brags about its social reach, client is the main element of the social network. A lot of research is being done on the problem of user classification in any social network. Identifying the most influential nodes is a significant issue in controlling the spreading cycle in complex networks. Centrality measures are utilized to rank the network nodes relying on different properties captured by that centrality. Researchers have been pursuing for quite a long time to outline a widespread strategy for user classification in a social network. In this project, we have attempted to devise a new technique, Improved Gravity centrality to group nodes in a complex social network utilizing the mix of network structure of the diagram and Gravity centrality, using H-index as the mass of the network node. We have thought about the aftereffect of our proposed strategy with different existing models in social organization writing on various constant datasets with the assistance of SIR epidemic model (Suspected-Infected-Recovered). When contrasted with existing techniques for node ranking, our outcomes provide quite an improvement.