Research On Mining Algorithms For Software Network Key Node Based On Complex Network

Mapping the complex software system to the model of complex software network, and then mining the key node that having important influence to the stability, reliability and security of the system is helpful to improve the stability and security of the software system, and has important theoretical significance and wide practical value and application prospect.In this paper, we take the function of the software system as the research object. The key function callers and key function callees are mined. The evolution process and the distribution characteristics of the relevant measurement values are also analyzed.Firstly, by combining the theory of complex network and the characteristics of complex software system, a approach of mapping complex software systems to directed software networks at function granularity is proposed. And then calculate the weight of each directed edge to present the close degree of dependence relationship.Secondly, based on the ability of accumulating vulnerabilities and the ability of propagating vulnerabilities of a node, we proposed a Software Network Key Node mi Ning approach(SN-KNN) based on local centrality. SN-KNN can mining the key caller nodes and key callee nodes respectively. Then, by considering the weight of directed edge, an improved approach KNMWSG is proposed.Thirdly, we give a definition of Directed Weighted Ripple Degree, which accord with the characteristics of complex software network. And based on the concept, we design a approach Alg_Imp RE to mine key nodes in directed weighted software network. We also conduct an analysis on the Saliency of Directed Weighted Ripple Degree, and the evolution law of Directed Weighted Ripple Degree is analyzed.Finally, empirical analysis on real software verifies the validity of our approach.