A Research Of Developer Recommendation In Open Source Software Community

With the pervasive application of the Internet,the open source software community is growing bigger and bigger.More and more developers participate in the software-related activities in open source community.Developers employ the open source community platform to collaborate with each other and exchange development experience in related area.The open source software community gradually accumulates a large number of data resources.In recent years,researchers have been aware of the value embodied in that kind of resources and attempt to utilize it to solve the problems in software engineering field with the technique of data mining and analysis.Depending on whether the recommendation context is related to a specific task,the problem of developer recommendation in open source software community can be classified into two categories.The first category is the task independent recommendation,whose main focus is ranking the developers in the community globally with the community's historical data resource.The rank list of this kind reflects a developer's contribution to the community in a specific aspect.The second category is the task dependent recommendation,whose main goal is finding out the most appropriate developers for a specific task during the software evolution process in an automated way.The solution of this problem helps to improve the developer's efficiency.One of the most active community is the bug tracking community,whose historical data resources reveal the law of developer's bug management activity.This paper carries out two aspects' research work.On one hand,the paper proposes one method to construct developer collaboration graph in the bug tracking community and use random walk model on that graph to rank the developers globally.The experiment result shows that developers who make great contribution in the community can be identified;on the other hand,this paper finds out that strong correlation relationship exists between bug report assignee and bug component with the technique of data analysis.With this finding we propose a bug assignee recommendation method based on bug component information.The comparative experiments demonstrate the effectiveness of this method.