摘要： 针对金融机构操纵风险具有构成复杂、涉及诸多复杂因素、难以结构化、缺少历史数据等特点，将贝叶斯网络技术引入银行操作风险建模。银行操作风险是由不完善的或有问题的内部程序、人员及系统或外部事件所造成损失的风险，难于建模与度量。贝叶斯网络是基于贝叶斯决策理论的因果建模技术，它很好地用于建立操作风险度量系统并作为操作风险度量的基础。通过实例演示了贝叶斯网络在银行操作风险方面的建模与应用，给出基于贝叶斯网络的银行操作风险管理的系统构架。

关键词: 操作风险, 贝叶斯网络, 因果建模, 情景分析, 不确定性推理

Abstract: This paper uses Bayesian network to model operational risk of bank that is difficult to model and measure. Operational risk is the risk resulting from inadequate or failed internal processes, people and systems or from external events. After reviewing operational risk of bank and Bayesian networks, this paper illustrates how to use Bayesian networks to manage operational risk by examples, including measure risk, casual modeling, scenario analysis and etc. This paper also gives a Bayesian network model framework of operational risk.

Key words: operational risk, Bayesian network, causalmodeling, scenario analysis, uncertainty reasoning

中图分类号: 

• TP319：F830