文章基本信息

标题：Visualization of Object Relations for Outlier Finding
本地全文：下载
作者：Takeo Okazaki
期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2014
卷号：14
期号：9
页码：5-11
出版社：International Journal of Computer Science and Network Security
摘要：The visualization of the relationship between objects such as distance and influence is one of the useful technique for outlier finding. Asymmetric data are particularly complicated. In this study, I proposed a method of two-dimensional coordinate arrangement according to dissimilarity between objects by decomposing the distortion and the symmetrical part. For the symmetrical part, I proposed an optimization method of deployment as a whole, which was an extension of the optimization techniques of conventional. For the distortion part, I defined the two components of the direction and magnitude to the skewness. The results of application experiments with two cases of real data showed the effectiveness of the outlier detection by the two-dimensional display of the distortion section in particular
关键词：2-dimensional visualization; Asymmetric relationship; Outlier