开源社区级联崩塌效应分析及基于SVM的项目失败预测  

Analysis of Cascade Collapse Effect of Open Source Community and Project Failure Prediction Based on Support Vector Machine

作　　者：张翔[1] 周健[1] ZHANG Xiang;ZHOU Jian(School of Computer Science and Technology,Anhui University,Hefei 230000,China)

机构地区：[1]安徽大学计算机科学与技术,合肥230000

出　　处：《小型微型计算机系统》2021年第5期1103-1108,共6页Journal of Chinese Computer Systems

基　　金：2018教育厅重点项目(Y06070799)资助;安徽省自然科学基金项目(J10118520150)资助;安徽省教育厅自然科学研究重点项目(J10118520444)资助.

摘　　要：以Github社区为例,通过采集海量社区项目的数据,分析了开源项目在开发过程中的风险传递和级联崩塌反应.通过重点分析技术关联和合作关联这两种开源项目之间最为常见的风险传递模式,结合采集数据,得出单一项目失败会产生一定规模的级联崩塌反应.其次,针对Github开源社区的大量成功和失败项目的数据,通过设计合理特征,基于支持向量机对成功与失败的项目数据进行训练,通过数据清洗和优化方法,使得训练得到的模型可以较好的对项目失败风险进行预测,对于开源社区的长久发展和风险评估提供了有效依据.Taking GitHub community as an example,this paper analyzes the risk transfer and cascading collapse response of open source projects in the development process by collecting massive community project data.Based on the analysis of the most common risk transfer mode between the two open-source projects,technical association and cooperative association,combined with the collection of data,it is concluded that a single project failure will produce a certain scale of cascading collapse response.Secondly,aiming at a large number of successful and failed project data of GitHub open source community,through the design of reasonable characteristics and support vector machine,the successful and failed project data are trained.Through data cleaning and optimization methods,the trained model can better predict the project failure risk,which provides an effective way for the long-term development and risk assessment of open source community basis.

关 键 词：开源社区 级联崩塌效应 支持向量机 分类预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]