Abstract: A software project risk management model based on Bayesian networks is presented. Using new project data obtained from the process of software development, it can continually predict risks, identify sources of risks, and take proper measure to reduce risk occurrence probability. Practice proves that the model is able to manage risks, and improve the success rate of software development.

Key words: Bayesian networks, Software project, Risk management model, Sources of risks

摘要： 提出了一种基于贝叶斯网络的软件项目风险管理模型。随着软件项目的进行，该风险管理模型能够利用不断更新的项目数据持续地预测潜在风险，确定风险源并采取适当的应对措施降低风险发生概率。经实践检验，在软件开发过程中引入该风险管理模型能够有效地对风险进行管理，提高软件开发的成功率。

关键词: 贝叶斯网络, 软件项目, 风险管理模型, 风险源