基于数学形态学的模糊异常点检测

异常点检测作为数据挖掘的一项重要任务,可能会导致意想不到的知识发现.但传统的异常点检测技术都忽略了数据的自然结构,即异常点与簇的联系.然而,把异常点得分和聚类方法结合起来有利于对异常点与簇的联系的研究.提出基于数学形态学的模糊异常点检测与分析,把数学形态学技术和基于连接的异常点检测方法集成到一个模糊模型中,从异常隶属度和模糊隶属度这两个方面来分析对象与簇集的模糊关系.通过充分的实验证明,该算法能够对复杂面状和变密度的数据集,正确、高效地找出异常点,同时发现与异常点相关联的簇信息,探索异常点与簇核的关联深度,对异常点本身的意义具有启发作用.

Fuzzy Outliers Detection Based Oil Mathematical Morphology

As one of the most important tasks in data mining,outliers detection may result in unexpected knowledge discovery.But the traditional technology ignored the natural data structure that is the relation between outliers and clusters.However,it is sometimes beneficial to integrate outlierness and a cluster method to enhance both outlier and cluster analysis.In this paper,a new algorithm of outlier detection named fuzzy outlier detection is presented based on mathematical morphology.which integrates mathematical morphology technology and connectivity-based outlier detection method in a fuzzy model.Thus,a pattern can be analyzed from two scopes:its outlierness and its cluster membership.The extensive experiment results illustrate that the proposed algorithm is an efficient outlier detection method that can discover all outliers effectively,accurately and precisely in the different databases.and explore the relation of an outlier to the clusters in order to discover the useful information hidden in the data set.