Rumor Source Detection on Interconnected Social Networks

Social networks are the most widely used platform for spreading information but it is also used to spread false rumor, using source identification we can held the source of spreading rumor responsible, also it can defeat the rumor as well, it has been the area of research for many researchers but due to limitation of different proposed study, it could not be used in the real environment, in this study we propose a methodology that is capable of identifying the rumor source on interconnected network. Interconnected network is considered as the network in which a rumor is propagated from one network to another and identifying the rumor from an independent network does not meet the requirement of identifying the real source. In this study we have proposed a methodology by which we can achieve high performance and accuracy based on an ensemble fusion and maximum voting on different centrality measures. We evaluate our model on two real datasets Facebook and U.S. Power Grid using error distance, this is the first attempt identifying the rumor source on interconnected network but for the acceptability of this model, we evaluate our results with the other approaches of rumor source identification on single network and found satisfactory results and our model outperform LPSI, we also compare our ensemble model with classical centrality based model ecc+clo which is most accurate model till date and found that our model outperforms the classical centrality’s model.