The User Influence Evaluation In Open Source Software Social Network Based On Interactive Information Extraction

In the era of data sharing and development,Open Source Software attracts more and more attention.Users and their interactions in the community constitute a complex social network.It is of great significance for information recommendation,community planning and coordination,and further system deployment to measure the user influence in Open Source Software Social Network and explore its spreading and action mechanism.The existing influence evaluation methods fail to highlight the characteristics of Open Source communities and make efficient use of all kinds of interactive information between users,which leads to low efficiency and unsatisfactory evaluation results.In view of this,based on the different interaction information of users,this paper studies the social influence evaluation of users in Open Source communities from the perspective of groups and individuals.Aiming at the problem of users' group influence evaluation,a method for nodes influence maximization in Open Source Software Social Network based on user spreading model is proposed.Firstly,users' follow information and records of their joint participation in projects in the community are extracted to establish directed networks.Next,quantifies the feedback of users on the project from three aspects(i.e.,approve,save and modify)and establishes a new probability propagation model.Then,based on the heuristic idea of influence maximization of nodes reverse ranking,a two-stage algorithm for evaluating user influence is proposed.Finally,the proposed theory and method are applied to Git Hub,a typical Open Source Software Social Network,and the correctness and effectiveness of this method are verified from influence spread and speed.Aiming at the problem of users' individual influence evaluation,a method for nodes influence ranking in Open Source Software Social Network based on the K-shell decomposition of weight-degree correlation is proposed.Firstly,users' cooperation information and interaction records are extracted,and three cooperative networks with different linguistic attributes are established.Next,the community is divided by Louvain algorithm.Then,the edge weight-degree correlation factor of the network is analyzed,which integrates the nodes' own attributes and neighbor roles,and updates the weights continuously during the K-shell decomposition process to obtain a new measure KWDC-shell.Finally,the proposed method is applied to Github platform.Experiments show that this method can distinguish nodes influence more accurately in networks with different linguistic attributes in terms of monotonicity and correctness,and has good performance.This paper integrates interdisciplinary knowledge and proposes a method for evaluating the social influence of users,which greatly improves the efficiency of the algorithm.It is conducive to further explore the community evolution law and collaboration mechanism of software ecosystem.Therefore,this paper has important research significance and value.There are 13 figures,4 tables and 87 references in this paper.