关键词: 产品伤害危机, 原因挖掘模型, 故障树, 贝叶斯网络

Abstract: In order to effectively solve the problem of improper selection and implementation of response measures caused by inaccurate disclosure of the causes of product quality crisis events, according to the discovering results based on the influencing factors of product quality crisis, the possible causes of product quality crisis events are extracted, and the relationship model of potential causes of quality crisis events is constructed by using fault tree method. Combining the fault tree method and the Bayesian network, based on expert investigation and fuzzy set theory, we infer the underlying cause prior probability algorithm, input the prior probability into the Bayesian network model to obtain the underlying cause posterior probability and critical importance degree, and use the posterior probability and critical importance degree together as the diagnosis basis of the quality crisis key causes. We build the cause mining model of product quality crisis event. The accuracy of the key causes mining results is verified through practical cases, which provides a theoretical basis for companies to mine the causes of quality crisis events.

Key words: product quality crisis, causes mining model, Fault Tree Analysis(FTA), Bayesian Network(BN)