Research On The Factors Influencing The Survival Status And Survival Prediction Of Open-Source Projects

With the rapid development of Open-Source thinking,more and more software development work is being done in the form of "open source".A large number of Open-Source projects gradually interdepend on and influence each other.Open-Source projects are increasingly linked as nodes into huge networks.But not all Open-Source projects in this network can develop healthily and stably in the long run.More Open-Source projects are affected by internal and external factors,and cannot develop healthily.In a tightly connected ecosystem network,when one project tends to die it can cause the whole network to collapse.To prevent such an event from occurring It is necessary to predict the survival of Open-Source projects in the ecosystem reasonably and effectively.To solve the above problems,the following work is carried out in this thesis: 1)This thesis proposes a bottom-up approach to analyze the influencing factors.The approach identifies a set of critical factors that affect the survival of open-source projects at both the project level and the ecosystem level.These factors provide the basis for research on the problem of predicting the survival of open-source projects.2)This thesis propose a hybrid structure method for predicting the survival of opensource projects.This method first designs a hybrid structure-based prediction model for the survival of open-source projects.Then this method constructs Graph Convolutional Networks(GCN)to characterize the ecosystem environment of open-source projects and further improve the prediction model performance.The project survival prediction model proposed in this thesis at the open-source ecosystem level performs well in evaluation metrics such as recall,accuracy,F-measure value,and AUC score,which is a significant improvement over traditional-models.And by comparison,it was found that ecosystem-level influencing factors can effectively improve the prediction of the model.In this thesis,we hope that this model can bring the following contributions: 1)The model can provide an important reference for developers in selecting suitable open-source projects for production activities;2)The model can provide an important basis for researchers to measure the health of open-source projects;3)The model can help managers of open-source projects in maintaining the health of their projects.