Based on local domain identifying influential nodes in social networks

Due to the boom of social networks, research on maximizing influence has become increasingly important and various disciplines are attracted by influence maximization. The key to maximizing impact is to identify a set of influential nodes that are widely distributed across the network. Therefore, to select the influential nodes of networks, we propose the LVoteRank algorithm. In LVoteRank, based on the theory of three-degree division, it is reasonable for nodes to gain different voting ability according to the different number of neighbors in 3-hop. In addition, to cut down the overlapping of influential regions of spreaders in networks, LVoteRank discounts the voting ability of neighbors in 3-hop of the selected nodes. Then, to reduce the number of update node voting scores, only recalculate the voting scores of neighbors in 4-hop of the selected nodes. To demonstrate the effectiveness of the LVoteRank, the Susceptible-Infected-Recovered and Linear Threshold models are simulated the spreading progress. Experimental results show that in most cases, LVoteRank is superior to the others methods in spreading speed and infection scale.