Research On Key Function Mining Algorithm Based On Software Characteristics

With the gradual popularization of software applications and the ever-increasing scale,software security issues have attracted the attention in all walks of life.In order to better understand the characteristics of the software network and its topology structure,from the perspective of data mining,important nodes in the software network are discovered,and then focus on important nodes in the network.It is of great significance for software fault detection,auxiliary software for software fault detection system maintenance,testing and prevention of bugs,and incorrect positioning.Based on the knowledge of the complex network and the characteristics of the software network,this paper researches functions in the software network.The main work is as follows.First of all,in this paper constructs a model for the undirected weighted software network and directed weighted software network.According to the static topology and dynamic calling features of the software network during the execution process.Using the Pvtrace software to track the data,understand the software execution path,and the software are executed multiple times.to chance,cleaning data integration effectively,has paved the way for the software network to identify the key function.Secondly,an important node discovery algorithm NDET based on evidence theory is proposed.The algorithm quantifies the importance degree of the function and puts forward the fluctuation value fv(i)based on the topological structure of the software.Finally,the node importance is measured by NI(Node Important).It provides a direction for analyzing static topology in software network.Thirdly,in order to solve the problem of the key function identification in dynamic software network.A new IKN algorithm for identifying key nodes is proposed.It not only considers the node communication ability from overall position of the nodes in the network but also measures the influencial ability of the function node from the own local characteristics of the nodes in the network to identifiy key nodes more accurately,which provides the directions for software failure prediction and system maintenance.Finally,the validity of the two proposed algorithms in this paper is verified,and the experimental results are compared and analyzed.By analyzing the evolution rules of different versions of software network,new ideas can be provided for further understanding of software networks.