Research And Application Of Software Network Node Influence Mining Algorithm Based On Complex Network

In recent years,with the rapid development of Internet,people's life has brought a lot of convenience,but also brought a series of network risks.In the era of Internet,the security and stability of software system is becoming more and more serious.This paper,based on complex network,combines the network topology of the software system's function call relationship analysis software system,and proposes a community discovery algorithm based on the node influence,and then analyzes the structure characteristics of the software system and provides help for the update of the software system.First,a new weighted network model is proposed based on the call relation between functions.According to the relevant information of function call in the execution process,and according to the proportion of the real calling sequence of the software as the weight,the DFS-WEIGHT algorithm is proposed to construct the network model.Secondly,based on the calling relation and calling order between functions,the influence of nodes on the whole network is analyzed,and the influence algorithm INM-SN is proposed.According to the influence channel in real life,the degree influence of other nodes to the current node is determined,and the degree influence of other nodes on the current node is used as the overall influence of the current node in the network.In the process of comprehensive analysis,the influence of nodes in the complex directional weighted software network is identified respectively.Thirdly,according to the influence model,we introduce the community theory in complex networks,and propose a community discovery algorithm based on the influence of nodes CDABNI.Firstly,according to the influence ranking of every node,the algorithm determines the seed node of the community and initializes the community.Then,according to the seed nodes of each association,according to the degree of interdependence and the degree of dependence on the community,the expansion strategy is adopted to divide the remaining nodes into the affiliated communities,and the optimal partition is obtained by iteration to get the final result of the community division.Finally,through the experiment of the two software,based on the real software,the distribution of the influence of the nodes in different versions of the software and the result of the community division are analyzed.The rules of the software in the process of updating are analyzed,and the correctness and validity of the proposed algorithm are verified at the same time.