Identifying influential nodes based on local dimension

How to identify influential nodes in complex networks is still an open issue. In this paper, we propose a novel method to identify influential nodes based on the local dimension (LD) of each node, where low LD values are suggestive of high influence. Applied to four real networks, our method has been demonstrated to have a comparable ability of identifying influential nodes with other commonly used methods. Furthermore, our method performs much better than the k-shell decomposition method, especially in the network with community structure. It can not only identify the influential nodes but also subdivide the nodes in the innermost layers.