With the popularity of online social network these have become important platforms for the spread of information. This not only includes correct and useful information, but also false information, and even rumors which could result in panic. Therefore, the containment of rumor spread in social networks is important. In this paper, we propose an effective method that involves selecting a set of nodes in (k, η)-cores and immunize these nodes for rumor containment. First, we study rumor influence propagation in social networks under the extended Independent Cascade (EIC) model, an extension of the classic Independent Cascade (IC) model. Then, we decompose a social network into subgraphs via core decomposition of uncertain graphs and compute the number of immune nodes in each subgraph. Further we greedily select nodes with the Maximum Marginal Covering Neighbors Set as immune nodes. Finally, we conduct experiments using real-world datasets to evaluate our method. Experimental results show the effectiveness of our method.
Keywords: (k, η) cores decomposition; containment of rumor spread; extended independent cascade model; greedy algorithm; social networks.