Identification of influential nodes in multilayer networks using the concept of community representatives

Selection of seed nodes (influential nodes) in socialnetworks is very important and necessary. Because these nodescan infiltrate other nodes and bring dynamics and control ofinformation dissemination in the network. The goal of influencemaximization (IM) is to select the minimum number of seed nodesto maximize the information diffusion or social impact. This issueis necessary for advertising, viral marketing, rumor control, etc.Many studies have been done on the influence maximizationproblem in single layer or monoplex networks. In multiplexnetworks where the nodes in each layer may have differentconnections with other nodes, identifying seed nodes for influencemaximization has a different nature, and the desired results withclassical methods may not be obtained. In this article, acommunity-based influence maximization method in multiplexnetworks is proposed. The basis of the proposed method is todetermine the representative Seed for each community. Therepresentative node is selected in a way that covers morecommunities. In the proposed method, first, multi-stage filtering isperformed on the nodes inside each layer to remove leaf nodes andnodes with low degrees. Then the communities inside each layerare identified with classical methods and the list of nodes of eachcommunity is maintained. After identifying the communities, theseed nodes for each community are selected according to thecomplete structure criteria of the layers of the multiplex network.The efficiency of the proposed method has been evaluated on realdatasets and the obtained results show high accuracy indetermining seed nodes.