Research And Implementation Of Open Source Software Recommendation Method Based On Crowd Participation

Improve the efficiency and quality of software development is a huge challenge in the field of software engineering,and software reuse is the fundamental way to solve this problem.With the rapid development of open source software,more and more software developers,amateur enthusiasts participate in the development of open source software,share their own experience knowledge,publish their own software projects,which has accumulated vast amounts of software resources in the open source software community.O n the one hand,the vast amount of software resources provides a rich,reusable resources for software reuse.On the other hand,Open source software has the characteristics of huge scale,wide distribution and rapid growth,which makes the open source software is facing a great challenge in the search,cross community resource location and correlation analysis.To address this challenge,this thesis has carried out these works:(1)We propose a recommendation algorithm based on software project labeling.In the open source community,tagging tags is a kind of important attribute of software project.It is the theme,function,characteristic and other macro aspects of the software,which has a good organizational structure.In this paper,a detailed analysis and discussion on the label of the software is carried out,and the recommendation algorithm based on item label is proposed,the algorithm is bas ed on a variant of the content recommendation,which can effectively reflect the similarity between the software through the characteristics of the software.By comparing the experimental results with the text description information,this algorithm has a good effect on the accuracy of the evaluation.(2)We propose a Cross-Community-Association of open source software recommendation algorithm.In the open source ecosystem different types of software resources are often distributed in different communities.Developers to develop and publish their own software projects in Collaborative-Development community,while the user is in the Knowledge-Sharing community comments,consulting related software issues.The data of the two types of communities are relatively scattered,there is no significant correlation,which gives the software defect repair and software resource location with a huge challenge.The recommendation method based on the project label tag,although it can ease the problem in the similar software,but the related software is less effective.In this paper,we propose a recommendation algorithm based on the resource characteristics of the two kinds of communities.In comparison with the recommendation algorithm based on tags,this algorithm is more prominent in the relevant software.(3)We designed and implemented the open source software recommendation system.In order to display the results of the experiment,this paper designs and implements the open source software recommendation system based on OSSEAN platform.At present,the system has generated a corresponding recommendation project for 231542 open source software projects,and has realized the cross community's software resources.