基于距离和的孤立点用户意义分析算法及应用.pdfVIP

重庆理工大学学报(自然科学) 2010 年 1 月 第 24 卷第1 期 Vo l. 24 No.1 loumal of Chon吨gq甲m吨19 Univ刊e臼ers毡8叫 of Techn叫olog剧y叭( Natural Science ) Jan. 2010 基于距离和的孤立点用户意义分析算法及应用 王越1 ， 2 刘亚辉徐传运2 ( 1.重庆邮电大学计算机科学与技术学院，重庆 400065;2. 重庆理工大学计算机软件及应用研究所，重庆 400050 ) 摘 要:有关孤立点分析的研究大多集中在如何高效地识别出孤立点，很少有研究提供对识 别出的孤立点用户意义的解释。针对这一情况，提出了一种基于距离和的孤立点用户意义分析 算法，针对每个孤立点都给出了其原因属性以及对该原因属性的描述，即给出了孤立点的用户意 义。最后将该算法应用到质量管理中进行了实验。实验结果表明:该算法是行之有效的，有一定 实用意义。 关键词:孤立点;用户意义;距离和;数据挖掘;质量管理 中图分类号:TP31 1. 13 文献标识码:A 文章编号:1674 -8425( 2010 )01 -0055 -05 Analysis Algorithm and Application of Distance Sum-based Outlier Customer Meaning WANG Yue 1, 2 , LIU Ya-hui 1 , XU Chuan-yun2 (1. College of Computer Science and Technology , Chongqing University of Posts and Telecommunications , Chongqing 400065 , China; 2. Institute of Software Technology and Application , Chongqing University of Technology , Chongqing 400050 , China) Abstract: The studies on outlier analysis mostly focus mostly on the identification aspect , rarely provide an explanation of the outlier customer meaning. The explanation of outlier is usually very important to users , and it is as important as the outlier in many applications. In this paper a new distance sum-based outlier customer meaning analysis algorithm is put forward , the reason auribute and its descriptions are given for every outlier in this algorithm. Then the algorithm is applied to quality management for experiment. The experimental results show that the algorithm is effective and practical. Key words: outlier; customer meaning; distance sum; data mining; quality management 孤立点分析是数据挖掘技术中一个非常重要 个子问题:①什么样的数据是不一致的，即孤立，点 的研究方向，它是从大量