A Node Importance Evaluation Method for Complex Networks Based on M-order Neighbour Importance Contribution

Abstract

Presently, most of the existing node importance evaluation algorithms cannot completely reflect the result of common effect among the nodes in the network. With this consideration, the conception of m order neighbor node is imported to propose a node importance evaluation method for complex networks, and two parameters are introduced to respectively adjust the dependency degree of node importance evaluation to own characteristic of nodes and m order neighbor nodes. Experiments use the topology of ARPA network to evaluate the node importance situation of algorithm in m value-taking conditions. The evaluation results show that, compared with some traditional evaluation methods, the suggested method has higher evaluation accuracy and efficiency; furthermore, the results also show an important dynamics phenomenon that when the investigated depth m value of neighbor node is greater than average path length L of the network, the suggested method can have a reliable evaluation result with fairly high accuracy.