Study Of Developer Value Metrics And Recommendation Algorithms Based On Social Technology Network

GitHub is a well-known open source code hosting platform.There are a large number of open source projects and developers involved in open source projects on GitHub,and developers have accumulated their own experience and developed their own expertise and value through rich collaborative behavior.With the development of the open source community,more and more developers are gradually stepping out of their comfort zone and starting to explore new areas of open source projects.When developers find a target project,how to quickly understand the core developers of the project and invest in the community is a key issue.Currently,only a partial list of developers can be viewed on the GitHub open source repository home page,but it is not possible to accurately determine the core developers of a project and thus collaborate with them.Therefore,this study uses the developer collaboration events on GitHub as the basis,builds a developer social technology network to measure developer value,uses the polarity characteristics of developer comment text information and the importance of nodes in the network,and solves the developer recommendation problem in the open source community based on the link prediction algorithm to help developers understand the core members of the project,and promotes developers to collaborate with high-value developers with similar collaboration patterns.The main contributions of this study are as follows:(1)Capture data and build developer social technology network:In order to implement developer recommendations on the GitHub platform,we first need to collect information about developers on the platform.Due to the limitation of the number of accesses to the API provided by GitHub itself,it is impossible to obtain relevant data directly from GitHub.Therefore,this study uses OpenDigger,a common data collection and processing framework for GitHub,to collect and process engineering data on the GitHub platform,to build a data processing system for GitHub,to analyze the quality of the collected data,and to build a developer social technology network using the behavioral data of project developers.(2)Propose a developer value metric model based on developer social technology network:Opensource communities have community attributes in their development,and the definition of value in sociology fits the network attributes in open source communities to a certain extent,so this study integrates the traditional developer capability assessment methods,the characteristics of open source communities and the relationship,structure and cognitive dimensions of social capital theory to propose a developer value metric model for open source communities——OpenCapital,and the authority of the model proposed in this study is verified by qualitative and quantitative analysis based on the developer social technology network.(3)Propose a developer recommendation algorithm based on open source collaboration:A link prediction algorithm is used to provide more accurate open source developer recommendations,thus helping developers to integrate into suitable open source projects quickly and efficiently.This study finds that developers’ collaborative communication is mainly concentrated in various comments such as Issue and Pull Request,so the traditional link prediction algorithm is combined with the textual information involved in developers’ collaborative behavior,and the sentiment analysis of textual content is used as an auxiliary feature for developer recommendation,and the importance of nodes in the network is also combined to propose a developer recommendation algorithm for open source collaboration LPOSC,and experimentally demonstrate the accuracy and effectiveness of the algorithm for making recommendations.In summary,this study uses a data collection and processing framework to obtain and store the required datasets for the developer recommendation scenario in the open source domain,which is more efficient and accurate than the APIs provided by GitHub itself,and builds a developer social technology network based on the acquired developer behavior data as the network foundation for developer recommendation.To address the problems of developer contribution and value measurement in the open source field,this study proposes a developer value measurement model for open source communities by combining the characteristics of open source communities and multiple dimensions of social capital theory,and experiments prove that the model can more accurately measure the value of developers.In addition,this study analyzes the textual information involved in the collaborative behavior of developers and proposes a better developer recommendation algorithm by combining the attributes of nodes in the network,and experiments prove that this algorithm can effectively improve the accuracy of recommendations.