Research On Nodes Importance Evaluation In Complex Networks

Complex networks have been a common hot research topic in a great number of fields,including nature science, engineering, economy, management and so on.As a matter of fact,we can find a lot of network instances either in human society or nature. To start anything in view of reality, a majority of artificial or natural systems can be denoted by network graphs composed of many interacted nodes,using the theory of complex networks,from different research angles.Evaluating nodes importance is an important research task in complex networks,which attracts so many researchers studying for it,since finding important nodes in complex networks has very crucial application value in too many fields, such as society, politics, medicine,information technology and so on.Nodes importance evaluation methods in complex networks are studied in this paper. We mainly have done some research in this paper as following.1.In this paper, we summarize nodes importance evaluation methods in complex networks from three aspects,that are social relation network,system science and web searching,and analyze their characteristics respectively.2.We propose two modified methods for evaluating nodes importance in complex networks,which are based on contributing degree of nodes similarity and field force among adjacency nodes,after studying the shortcomings of Pagerank algorithm.The two methods suggest using the contributing degree of nodes similarity and field force among adjacency nodes respectively to reconstruct the modified probability transfer matrix among all adjacency nodes,and replacing uniform parameter in Pagerank algorithm by normed nodes closeness values,to reappraise nodes importance in complex networks.Experiment results indicate that the two methods not only can overcome the limitations of degree ranking, closeness ranking and betweenness ranking,but also can work out the limitation of treating all adjacency nodes equally in Pagerank algorithm,thereby, they can find important nodes in complex networks more precisely and effectively.3.In allusion to the shortage of BA model,the paper also gives two extended BA models using the above two proposed nodes importance evaluation methods,that are, nodes importance evaluation methods based on contributing degree of nodes similarity and field theory in physics respectively, and rewriting the formula of nodes preferential link mechanism.The two new extended BA models not only accord with power-law distribution, but also are better than BA model in three assessing guidelines,average shortest path length,clustering coefficient and average degrees of network,consequently, obtaining two more practical network evolvement models.