摘要
【目的】针对因果图传统推理方法中不交化割集的推理过程太过繁琐的问题,提出利用二元决策图(BDD)的推理方法。【方法】首先将因果图转化为标准化因果树,再转化为BDD结构,然后通过搜索路径直接得到简化的不交化割集。【结果】理论分析表明,该方法避开了求解最终割集的过程以及去除冗余的不交化割集的过程,从而降低了因果图推理中的计算复杂度。【结论】将新方法应用于银行外部欺诈风险实例中进行分析,通过与传统推理过程相比较,新方法更高效。
[Purposes]The inference process to use the traditional method of causality diagram to calculate disjoint cut sets is too cumbersome,so the reasoning method using binary decision diagram（BDD）is proposed.[Methods]Firstly,the causality diagram is transformed into a standardized causality tree,and then into BDD structure.After that,those disjoint cut sets which have been simplified are obtained by searching path.[Findings]The theoretical analysis shows that the method avoids the process of solving the final cut set as well as removing redundant disjoint cut sets,thereby reducing computational complexity in causality diagram inference.[Conclusions]Finally,applied in instance of Bank external fraud risk,by compared with the traditional reasoning,this method is more efficient.
作者
蒋杰
王洪春
JIANG Jie;WANG Hongchun(College of Mathematics,Chongqing Normal University,Chongqing 401331,China)
出处
《重庆师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2018年第6期56-63,共8页
Journal of Chongqing Normal University:Natural Science
基金
国家社会科学基金(No.13BTJ008)
重庆师范大学研究生科研创新项目(No.YKC17002)
关键词
因果图
二元决策图
标准化因果树
不交化割集
causality diagram
binary decision diagram
standardized causality diagram
disjoint cut sets
