基于概念层次树的例外挖掘方法研究

Research on Mining Outliers Methods on the Basis of the Concept Hi erarch y Tree

例外检测能够在诸如电子商务、信用卡欺骗及气象数据分析等领域中挖掘真正未预料到的知识,人们注意到,已有发现例外的方法只能处理数值型属性,同时不允许用户动态地改变参数。在此,文章基于“概念层次树”、“语义距离”及“语义贴近度”等概念,研究了SDB—例外及FCB—例外,并给出了挖掘这些例外的有效算法。

The identification of outliers can lead to the discovery of truly unexpected knowledge in areas such as elec-tronic commence,credit card fraud,and even t he analysis of meteorological data.Existing methods that people have seen for f inding outliners can only deal with numeric data attribute,and have no incremen tal techniques that allow the user to freely change the parameters.Here,on the basis of'concept hierarchy tree','the Semantic Distance'and'the Semantic prox-imity',this paper studies the notion of SDB-(Semantic Distance-Based Outliers)outliers and FCB-(Fuzzy Cluster-Based Outliers)outliers.Algorith ms for computing such outliers have been developed.