Research On Important Node Mining Method Of Software Function Association Network

Only a few nodes in the software network have important influence on the stability,security and reliability of the network.How to find valuable information in a large number of software data and discover nodes that have an important impact on the network is of great significance to ensure the development quality of software system,and it is also important for software security.This paper takes the function in the software system as a granular unit,combines the structural characteristics of the software network to study the propagation and accumulation of faults,mines the important nodes in the software system,and analyzes the evolution rules of the relevant metrics.The main work is as follows.Firstly,the structural characteristics and function execution trace of the complex software system are analyzed,from the aspects of static construction and dynamic execution of software.The network model of the complex software system is constructed by abstracting the function in the system as a node and analyzing the dependencies and dependencies between the functions.Secondly,the static topology of the software is analyzed,and an important node in static network mining algorithm based on structural entropy INMSE is proposed.Introduce the concept of sub-networks and it refines the topology of the network.Analyze the cumulative process of faults in the network and quantitatively calculate the impact of the call object in each network on its root node.Then,the fault propagation process in the network is further studied,and the important nodes that are prone to failure and easily propagated in the network are discovered.Thirdly,the dynamic execution trace of the software is analyzed,and the important node in the dynamic network mining algorithm based on the probabilistic ripple effect MIN-PRE is proposed.It is found the failure propagation tendency path,considering the dependencies between nodes and their closeness.On this set of paths,the propagation and accumulation characteristics of the faults in the network are analyzed,the quantitative calculation of the importance degree of each node is realized,and the important nodes that are easily affected in the network are discovered.Finally,the algorithm proposed in this paper is programmed in Java and c++ language under Windows platform.The effectiveness of the method and the operating efficiency of the algorithm are verified by experiments,and the evolution of the metrics of important nodes is analyzed.