基于粒度计算的分类属性数据离群点检测算法

针对基于距离的离群检测算法无法有效应用于分类属性数据集,本文提出一种基于粒度计算理论的对象离群程度计算公式。基于该公式所计算的对象的离群因子值,对所有对象进行排序,将排序后的前k个对象声明为离群点。为了使用相对简单的方法从分类属性数据集中查找离群点,文中构造了一个算法ODAGr C(Outlier detection algorithm based on granular computing)。理论分析和应用实例证明了ODAGr C算法的有效性和可行性。

Outlier Detection Algorithm for Categorical Data Based on Granular Computing

Many distance-based outlier detection algorithms were proposed in the past,which can not effectively deal with categorical data set. In this paper,a novel formulation is proposed for the outlier degree of objects that is based on the granular computing theory. Each object on the basis of its Outlier Factor is ranked and we declare the top k objects in this ranking to be outliers. In order to develop relatively straightforward solutions to finding outliers from categorical data set,an algorithm is constructed,named ODAGr C（ Outlier detection algorithm based on granular computing）. Theory analysis and example calculation both manifest that the ODAGr C is efficient and feasible.