Identification of Key Nodes in Complex Networks

Abstract

Identification of key nodes in network is of great significance in practical application. Some of the existing importance evaluation indexes have the defects of limited scope of application and incomplete evaluation results. Considering that the importance of nodes in complex networks is not only affected by a single factor, this paper proposes a key node identification algorithm based on multi-attribute weighted fusion (KNIA-MWF). We quantitatively analyze the influence of different attributes on nodes and assign the index weights by the combination of objective entropy weight method (EWM) and subjective analytic hierarchy process (AHP) to obtain the final topological importance (FTI) of each node which represents its structural importance. This algorithm can not only be used to identify the key nodes of different types of complex networks, but also be easy to be extended. Experimental results of removing key nodes in five kinds of networks show that KNIA-MWF can accurately identify the critical nodes in complex networks.