基于聚类融合的异常检测算法  

An outlier detection algorithm based on clustering ensemble

作　　者：苏晓珂[1] 王秉政[1] 

机构地区：[1]郑州轻工业学院计算机与通信工程学院,河南郑州450002

出　　处：《郑州轻工业学院学报（自然科学版）》2011年第3期8-11,共4页Journal of Zhengzhou University of Light Industry:Natural Science

基　　金：河南省科技攻关项目(092102210108);河南省教育厅自然科学基础研究计划项目(2010A520033);郑州轻工业学院博士科研基金资助项目

摘　　要：针对任意形状聚类算法用于异常检测时参数设置困难的问题,提出一种基于聚类融合的异常检测算法:设置不同的半径阈值进行多次聚类,统计每次聚类中标记为异常的簇频率,将频率高的簇作为真正的异常.在UCI数据集上对该算法进行实验,结果表明:本算法可降低直接将小簇作为异常的高误报率,并且能提供给用户更为友好的操作.An outlier mining algorithm based on the clustering ensemble was presented in order to reduce the reliance for users and decrease the high false positive rate due to taking the small size clusters as the outliers directly.Outliers can be found according to the abnormal frequency of every record.The algorithm is able to provide the user a more friendly operation.The experimental results on the real-life datasets showed that the proposed algorithms are feasible and effective comparing with other classical algorithms and can be used for mixed dataset.

关 键 词：异常检测 聚类融合 异常簇 任意形状聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]