Research On The Algorithm Of Indentifying Influential Software Nodes Based On Complex Networks

For software system stability and robustness of the problem,this article is based on the software in the network the invocation of the relationship between nodes,introduce the knowledge of the complex network theory,from the perspective of network topology,for complex software in a network recognition algorithm of node was analyzed,and influence and node in the network structure and analysis the influencing factors of sexual.First of all,this paper analyzes the statistical characteristics of complex network,believe degree,degree distribution,average path length in identifying influential nodes play an important role,and summarizes the influence of node identification algorithm for these algorithms are analyzed and compared,on the basis of this proposed by granularity function,the software system mapping method for complex networks.Secondly,based on Page Rank algorithm,proposed in view of the software function calls the relationship between the influence of the node identification shall not be entitled to algorithm,this algorithm calculation function node strength value,the influence of measurement is higher,the greater the influence of the node,then excavated nodes in a network of influence.In order to solve the PageRank algorithm in recognition of the influence in disconnected network node sorting is not the only drawback,the algorithm in the original network node plus a root node,the root node and all the original node in the network two-way connection,the new network is strongly connected network,and solves the influence node sorting is not the only problem.Again,according to the characteristics and functions of the software function call relations,proposed in view of the weighted network recognition influence of the node of the algorithm.The algorithm based on Kendall 's correlation coefficient of node in and out of the degree of the influence of the power function value and the nodes of the relations are discussed in this paper.And applies the algorithm to the software evolution,analysis the characteristic of software evolution potential.Finally,starting from the real experimental data,analyze the two kinds of algorithm is proposed in this paper.From the experimental results,the analysis and node in thenetwork structure influence of constraint.Through the two open source software for experiments,the influence of the mining software system node,verify the correctness and effectiveness of the proposed method.Analysis of experimental results influence the value of each node distribution,and node affect sexual factors.