Enhancing developer recommendation with supplementary information via mining historical commits

Xiaobing Sun, Hui Yang, Xin Xia, Bin Li

    Research output: Contribution to journalArticleResearchpeer-review

    29 Citations (Scopus)


    Given a software issue request, one important activity is to recommend suitable developers to resolve it. A number of approaches have been proposed on developer recommendation. These developer recommendation techniques tend to recommend experienced developers, i.e., the more experienced a developer is, the more possible he/she is recommended. However, if the experienced developers are hectic, the junior developers may be employed to finish the incoming issue. But they may have difficulty in these tasks for lack of development experience. In this article, we propose an approach, EDR_SI, to enhance developer recommendation by considering their expertise and developing habits. Furthermore, EDR_SI also provides the personalized supplementary information for developers to use, such as personalized source code files, developer network and source-code change history. An empirical study on five open source subjects is conducted to evaluate the effectiveness of EDR_SI. In our study, EDR_SI is also compared with the state-of-art developer recommendation techniques, iMacPro, Location and ABA-Time-tf-idf, to evaluate the effectiveness of developer recommendation, and the results show that EDR_SI can not only improve the accuracy of developer recommendation, but also effectively provide useful supplementary information for them to use when they implement the incoming issue requests.

    Original languageEnglish
    Pages (from-to)355-368
    Number of pages14
    JournalJournal of Systems and Software
    Publication statusPublished - 1 Dec 2017


    • Bug assignment
    • Collaborative topic modeling
    • Commit repository
    • Developer recommendation
    • Personalized recommendation
    • Supplementary information recommendation

    Cite this