基于危化品事故案例的决策规则提取算法研究*

为充分解析文本形式的危化品事故案例，探讨事故类型与大量事故致因间的因果关联，从化工行业入手，通过扎根理论对175起危化品事故案例进行因素提取，共确定5个子系统、19个因素、80个子因素，提出基于事故生命周期的事故致因框架；在此基础上，提出改进的Gra Apriori算法，以解决经典的Apriori无法考虑属性类别关系的问题，最终获得14个以事故类型为决策属性的规则。结果表明:这些规则从不同事故阶段细致地剖析了危化品事故所遵循的规律，重现事故的演变过程，为危化品事故风险管理提供准确可靠的预防策略和决策支持。

Extraction of decision rules for hazardous chemicals cases based on improved Gra-Apriori algorithm

In order to fully analyze the accident cases of hazardous chemicals in text form and explore the causal relationship between accident types and large amount of accident causes,starting from the chemical industry,175 accident cases of hazardous chemicals were extracted through the grounded theory,then a total of 5 subsystems,19 factors and 80 sub factors were identified,and an accident cause framework based on the accident life cycle was proposed.On this basis,an improved Gra Apriori algorithm was put forward to solve the problem that the classic Apriori could not consider the relationship between attribute categories,and 14 rules with the accident type as the decision attribute were obtained.The results showed that these rules analyzed the laws followed by hazardous chemicals accidents from different accident stages in detail and represented the evolution process of accidents,which provide accurate and reliable prevention strategies and decision support for the risk management of hazardous chemicals accidents.