Code Reviewer Intelligent Prediction in Open Source Industrial Software Project

作     者：Zhifang Liao Bolin Zhang Xuechun Huang Song Yu Yan Zhang 

作者机构：School of Computer Science and EngineeringCentral South UniversityChangsha410083China Department of ComputingSchool of ComputingEngineering and Built EnvironmentGlasgow Caledonian UniversityGlasgowG40BAUK 

出 版 物：《Computer Modeling in Engineering & Sciences》 (工程与科学中的计算机建模（英文）)

年 卷 期：2023年第137卷第10期

页      面：687-704页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：support of National Social Science Fund(NSSF)under Grant(No.22BTQ033) 

主　　题：Open source software pull request random forest knowledge graph 

摘      要：Currently, open-source software is gradually being integrated into industrial software, while industry protocolsin industrial software are also gradually transferred to open-source community development. Industrial protocolstandardization organizations are confronted with fragmented and numerous code PR (Pull Request) and informalproposals, and differentworkflowswill lead to increased operating costs. The open-source community maintenanceteam needs software that is more intelligent to guide the identification and classification of these issues. To solvethe above problems, this paper proposes a PR review prediction model based on multi-dimensional features. Weextract 43 features of PR and divide them into five dimensions: contributor, reviewer, software project, PR, andsocial network of developers. The model integrates the above five-dimensional features, and a prediction model isbuilt based on a Random Forest Classifier to predict the review results of PR. On the other hand, to improve thequality of rejected PRs, we focus on problems raised in the review process and review comments of similar *** a PR revision recommendation model based on the PR review knowledge graph. Entity information andrelationships between entities are extracted from text and code information of PRs, historical review comments,and related issues. PR revisions will be recommended to code contributors by graph-based similarity *** experimental results illustrate that the above twomodels are effective and robust in PR review result predictionand PR revision recommendation.