In recent years,the world's largest code hosting platform GitHub is making rapid de-velopment,which is dependent on the group participation and contribution.Behind such behavior,there is a huge number of open source data.The researchers use data to assess developers' contribution,which has become a hot research topic.Through the assessment of the contributors,A lot of technology has emerged,such as projects recommendation,reviewers recommendation and bug trail.Therefore,based on the GitHub open source community,this article focuses on col-laborative development,capability assessment and group recommendation instead of rec-ommending individual developers.The main work and contributions are summarized as follows:First,a personalized open source data collection technology is proposed,which al-lows researchers to select different collection directions and personalized data fields ac-cording to research needs.Secondly,this paper proposes a multi-dimensional assessment method of the developers capability.Multi-dimensional reflects on the combination of multiple primary classification and secondary indicators measure.Among them,the pri-mary classification is technical ability,management ability and cooperation ability respec-tively.Specifically,the secondary indicators of technical ability are efficiency and qual-ity.The secondary indicators of management ability are version release,branch merge,integration test and document management.The secondary indicators of collaboration are friendliness and communication.Finally,this paper put forward a method to find and recommend potential development group in the GitHub community.The method based on the existing groups'characteristics like frequent activities and outstanding capability,establishes a way to find potential developer group by setting threshold value.Then it combines text matching and capability to recommend groups. |