Research On Open-source Software Quality Evaluation And Recommendation Technology Based On A Graph Model

With the rapid development of open source software(OSS),more and more people begin to pay close attention to OSS.They take part in the process of developing,sharing and learning knowledge of OSS.The number of open source software increases exponentially.The huge number of open source resources gives users a large more choice,but also brings more challenges.Firstly,it is hard for users to evaluate the quality of OSS.Although there are millions of OSS,the proportion of high quality software is quite small.Secondly,these massive open source resources are of large amount,various types and highly distributed,which results in great challenge for developer to accurately and quickly address their desired resources.What's more,the prosperity of open source software has driven the construction of open source community.There exists a large number of open source communities.Due to the different functions,the open source communities can be divided into two categories: software development community that can provide abundant repository data and development process reports,and knowledge sharing community that contains a large amount of discussion about OSS.However,different communities have strong connections and interactions.Most of the current research about OSS care the connections between software,and few of them consider the relationship across communities from an overall perspective,and fail to cover the users and their activities,which should be an indispensable part of the OSS.This paper model the OSS communities as a graph,which combines different types of OSS data and their connections.Based on this graphic model,we first analyze the evaluation of OSS quality.Using different elements especially the connections between different communities to explore the correlation of OSS quality.Then based on the graph model and quality analysis,we build a recommendation system to recommend relevant software for user.Finally,we evaluate our recommendation result with other communities' recommendation result,and achieve a well result.