The Research Of Influence Maximization Based On Overlapping Communities And TOPSIS

Due to the rapid development of Internet technology in recent years,online social network applications have emerged,such as Sina Weibo,Facebook,Wechat,Twitter and so on.Massive data which has impact on people’s behavior will be produced when people use these applications.If these massive data are processed and used,it can produce great benefits.So the problem of influence maximization has become a new research hotspot.Most of the existing thesis ignore the overlapping community characteristics of social networks when they study the influence maximization problem;In addition,when selecting seed nodes,the initial seed set is not optimal due to the incomplete index of selecting influence nodes.In order to solve the above problems,this thesis proposes a multi label propagation algorithm based on node similarity in attribute network to detect overlapping communities and an influence maximization algorithm based on overlapping communities and TOPSIS to find the optimal initial seed node set.The main innovations of this paper are as follows:1.In order to solve the problem of ignoring the overlapping community characteristics of social networks in the research of influence maximization,a multi label propagation algorithm based on node similarity is proposed to detect overlapping communities.In the process of label propagation,the algorithm considers the network topology information and node attributes,calculates the similarity between nodes according to the network topology information and multiple attributes of nodes(labels)and takes them as the weight of edges.The experimental results show that the proposed algorithm has higher NMI and EQ than the classical algorithm.2.In view of the fact that most researchers do not consider overlapping communities and the index is not comprehensive when selecting influential nodes,a hybrid algorithm of multi label propagation algorithm and TOPSIS algorithm based on node similarity is proposed.This algorithm proposes a new index IOC(influence of overlapping community)to evaluate the influence of overlapping communities;A new index ION(influence of node)is proposed to evaluate the influence of nodes.Firstly,the IOC of the overlapping community is calculated,then the ION of the node is calculated.Then,the TOPSIS is used to analyze the two new indicators IOC and ION,as well as BC,DC,CC and Page Rank,to find the most influential seed set in the overlapping community.The comparative experiments on three public datasets show that the number of nodes affected by the initial seed node is more under the independent cascade information propagation model.