摘要
当监控系统或识别系统得到的数据值异常时,为了分析数据值异常的原因,发现异常的规律,基于模糊统计提出了一种自动判定数据属于一般或特殊状态的方法,并基于贝叶斯网(BN)挖掘数据中一般-特殊知识的各影响因素之间的相互依赖关系,给出一种发现关键影响因素的方法。实验表明该方法具有一定的可行性及合理性。
In real applications,when the data obtained from monitoring or recognition systems are abnormal,it is necessary to explore the inherent reason that data are abnormal on the observed objects.A fuzzy-statistics-based method was proposed for automatic recognition of the normal or abnormal states on given data sets.Then the method was proposed based on Bayesian Network(BN) to obtain the general and specific knowledge implied in data as well as the dependencies among relevant factors.In addition,the method for discovering the critical influence factors was given.Experimental results show the feasibility and rational of the methods.
出处
《计算机应用》
CSCD
北大核心
2008年第S2期213-216,共4页
journal of Computer Applications
关键词
依赖关系
模糊统计
贝叶斯网
关键影响
dependencies
fuzzy statistics
Bayesian Network(BN)
crucial influences