Community Detection Of Software Ecosystem Based On Complex Network

Software ecosystem can be regarded as a kind of social network with complex structure,which is the highest level in the field of software engineering.It has great theoretical and practical significance to explore the community structure in software ecosystem,such as project recommendation,impact assessment,and cooperation forecasting.At present,the community detection algorithm for traditional complex networks has been widely studied.However,the existing community detection algorithms do not fully consider the characteristics of a large number of information interaction and complex social relations in the software ecosystem,which makes the algorithm inefficient and difficult to obtain satisfactory community division results.Therefore,it is of necessary to study the theory and method of pertinence and detect the community of software ecosystem.In view of this,this paper focuses on the different social relationships between developers in the software ecosystem to build different types of networks,and studies community detection method of the software ecosystem.According to the cooperation between developers,a method of community detection in software ecosystem by comprehensively evaluating developer cooperation intensity is proposed.Firstly,combine network topology information and developer interaction information to calculate the developer cooperation intensity,so as to comprehensively evaluate the developers cooperation intensity from both topological and semantic properties.Then,a community detection algorithm,ABDCI is proposed based on the cooperation intensity of developers by referring to the hierarchical clustering idea of Louvain algorithm.Finally,this method is applied to developer networks in the software ecosystem through GitHub hosting platform.The experimental results find that this method can identify a clearer community structure for the developer collaboration network in software ecosystem.According to the concern relationship between developers,a method of community detection in software ecosystem based on comprehensive impact assessment of developers is proposed.Firstly,the information propagation gain of nodes is calculated by establishing a two-step information propagation model.Then,the contribution attribute information of nodes is used for non-dominant ranking,and combining the information propagation gain of nodes to sort the comprehensive influence of nodes.Then,community detection is carried out with the influential nodes as the center and the information dissemination probability as the clustering direction.Finally,collect data through GitHub API for experiment.The experimental results show that the proposed method has a good performance in both ranking of developer influence and the division of community structure.This paper studies the theory and method of community detection of software ecosystem from the aspects of domain knowledge extraction,network construction,model building,algorithm solving,etc.,which greatly improves the accuracy and efficiency of community detection,and lays a theoretical foundation for building a healthy and sustainable software ecosystem.Therefore,it has important theoretical significance and application value.There are 15 figures,10 tables and 84 references in this paper.