RESEARCH ON CASES KNOWLEDGE DISCOVERY BASED ON BAYESIAN NETWORK REASONING
Nowadays, the large amount of data has been described as data rich and information poor, and due to ever-changing world and limited understanding of human beings, the knowledge often contains some uncertain ingredient.Meanwhile, the quality and safety standards for commodities are essential to the enhancement of their quality and safety, and to the protection of the health and safety of consumers.MEBN is a hot research field in recent years, which is a new method of knowledge representation and reasoning based on Bayesian mechanism.With the combination of fault tree analysis, we can get the key factors set better, and provides decision support to improve and develop national standard.This paper presents a parameter learning algorithm based on D-S evidence theory concerning the partial data incompleteness, and deduces the key factors set by cluster analysis of posterior probability and four importance degree measures.Validated from a practical project which exploited the potential of these methods, the effectiveness of these methods was demonstrated in the safety standards building process of CNIS.This paper will have a certain reference value for government departments to identify key factor set of commodities.
Multiple Entity Bayesian Net D-S Evidence Theory Parameter Learning Algorithm Knowledge Discovery
Yi Hao Siyuan Wang Li Wang
School of Economics and Management, Beihang University,Beijing 100191, China
国际会议
The 11th International Conference on Industrial Management(第十一届工业管理国际会议)
日本
英文
503-508
2012-08-29（万方平台首次上网日期，不代表论文的发表时间）