Enhancement of Voting Scores with Multiple Attributes Based on VoteRank++ to Identify Influential Nodes in Social Networks

With the prosperity of social networks, Influence maximization is a crucial analysis drawback within the field of network science due to its business value. In this regard, we propose the EAVoteRank++, inspired by VoteRank++, to iteratively select the influential node. It is commonly recognized that degree is a well-known centrality metric for identifying prominent nodes, and neighbors’ contributions should also be considered. Furthermore, topological connections between neighbors have an impact on node spreading ability; the more connections between neighbors, the higher the risk of infection. Therefore, EAVoterank++ algorithm identify nodes’s voting ability by considering degree, position in network by improve k-shell decomposition and clustering coefficient as well as neighbors. The weights of attribute are calculated by entropy technology. Furthermore, based on VoteRank++, EAVoteRank++ minimizes the voting ability of 2-hop neighbors of the selected nodes to decrease the overlapping of influential regions of spreaders. To demonstrate the effectiveness of the proposed method, we employ both the SIR model and LS model to simulate the spreading progress, then calculate the accuracy of the proposed algorithm, compare with other methods. The experimental results with 2 propagation simulation models on 6 datasets demonstrate the good performance of the proposed method on discrimination capability and accuracy.