基于蚁群聚类算法的离群挖掘方法  

Research on Fault Diagnosis Based on Outlier Mining

作　　者：杨欣斌[1] 孙京诰[1] 黄道[1] 

机构地区：[1]华东理工大学信息学院,上海200237

出　　处：《计算机工程与应用》2003年第9期12-13,37,共3页Computer Engineering and Applications

摘　　要：离群挖掘是数据挖掘研究的重要内容,在实际生活中获得广泛应用。该文首先给出了离群数据的量化定义,并用基于蚁群的聚类学习方法,产生了状态空间的整体特征。然后结合具体的设备对象,提出了离群数据的挖掘方法。最后进行了实验验证,结果表明该文提出的方法是有效的。Outlier mining is an important issue in data mining community.It was applied widely in practice.A quantitative definition is proposed at the beginning of this paper.The integral character of the state space is created using the clustering method based on ant colony.And a method of mining outlier is put forward according to the practical device.At the end,an experiment has been done and the result proves the effectiveness of the method.

关 键 词：离群数据 数据挖掘 蚁群 聚类 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]