基于K-means的数据流离群点检测算法  

Data stream outlier detection algorithm based on K-means

作　　者：韩崇[1] 袁颖珊 梅焘[2] 耿慧玲 HAN Chong;YUAN Yingshan;MEI Tao;GENG Huiling(College of Computer, Nanjing University of Posts and Telecommunications, Nanjing 210003, China;College of Tongda, Nanjing University of Posts and Telecommunications, Yangzhou, Jiangsu 225127, China)

机构地区：[1]南京邮电大学计算机学院,南京210003 [2]南京邮电大学通达学院,江苏扬州225127

出　　处：《计算机工程与应用》2017年第3期58-63,共6页Computer Engineering and Applications

基　　金：江苏省自然科学基金(No.BK20150868);江苏省交通运输与安全保障重点实验室开放课题(No.TTS201502);江苏省无线传感网高技术研究重点实验室开放课题(No.WSNLBZY201505);南京邮电大学引进人才科研启动基金(No.NY214013);南京邮电大学通达学院大学生科技创新训练计划STITP(No.201513989010Y)

摘　　要：针对数据流中离群点挖掘问题,在K-means聚类算法基础上,提出了基于距离的准则进行数据间离群点判断的离群点检测DOKM算法。根据数据流概念漂移检测结果来自适应地调整滑动窗口大小,从而实现对数据流的离群点检测,与其他离群点算法的一系列实验验证和对比结果表明,DOKM算法在人工数据集和真实数据集中均可以实现对离群点的有效检测。Aiming at the problem of outlier mining in data streams,an outlier detection algorithm,DOKM algorithm,based on the K-means clustering algorithm,with distance-based criterion to evaluate the issues of outliers is proposed inthis paper.For making the detection result of the concept drift of data stream it automatically adjusts the sliding windowsize and detects outliers in the data stream.A set of experiments show that the proposed DOKM algorithm can achieve theeffective detection of outliers in both an artificial data set and a real data set compared with other outlier detection algorithms.

关 键 词：概念漂移 数据流 K-MEANS聚类 可变滑动窗口 离群点检测 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]