Global recursive based node importance evaluation

Abstract

The world city network (WCN) research has been promoting the evaluation algorithms in complex networks. The study of urban network focuses on the measurements of city position in the WCN. In previous study, a set of algorithms such as centricity evaluation, power evaluation and their recursive power are proposed to support the WCN research. In this paper, we propose a novel global recursive based node importance evaluation (GRNIE) algorithm for WCN, which improves the performance in evaluating the network centricity and power by being applied to the Friedmann Basics Network. The results of the experiment show that the proposed GRNIE outperforms the previous algorithms (i.e., degree, recursive centricity, recursive power) in task of classification with the improvements of 72 %, 32 % and 20 % respectively in accuracy.