Ranking the spreading influence of nodes in weighted networks by combining node2vec and weighted K-Shell decomposition

Controlling the spreading process in complex networks is of great theoretical and practical significance. One of the important methods to control the spreading process is ranking the spreading influence of nodes. However, it is still an opening topic to perform an accurate method to rank the spreading influence of nodes on weighted networks. Inspired by the idea that the similarity of topological structure of a node and its neighbours can be utilized to effectively rank the spreading influence of nodes, we propose a novel method termed as node2vec weighted K-Shell (N2VWKS) centrality. First, a graph embedding method node2vec is introduced to calculate the similarity of topological structure between nodes. Then, the weighted K-Shell decomposition method (WKS) is utilized to consider the weights of edges and capture global topological structure information. Further, N2VWKS is proposed by combining the similarity between nodes and the effect of weights. We evaluate the ranking performance by utilizing two common metrics, namely accuracy and resolution. Experimental results show that N2VWKS exposes better performance than other methods.