文章基本信息

标题：Linkage Pattern Mining using Interval and Order of Pattern Appearance
本地全文：下载
作者：Saerom Lee ; Kaiji Sugimoto ; Yoshifumi Okada 等
期刊名称：IAENG International Journal of Computer Science
印刷版ISSN：1819-656X
电子版ISSN：1819-9224
出版年度：2019
卷号：46
期号：4
页码：691-698
出版社：IAENG - International Association of Engineers
摘要：Linkage pattern mining is a data mining technique employed to extract a linkage pattern, that is, a set of frequent patterns appearing repeatedly across multiple sequential data. In this technique, when frequent patterns appear in the same time zone for multiple sequential data, they are extracted as a linkage pattern even if these patterns are neither correlated nor similar. Thus, this mining method is expected to become a promising approach for predicting the risks associated with disease and analysis of voice data. However, the existing linkage pattern mining method cannot extract the linkage pattern in which no overlap on the time axes is identified in its frequent patterns even if those frequent patterns obviously show a continuous appearance. In addition, there is another serious problem in the existing method; namely, in any two linkage patterns composed of the same frequent patterns, even if the order of frequent patterns for each other is different, these linkage patterns are mistakenly regarded as an identical linkage pattern. To solve these problems, we propose a new linkage pattern mining method that considers the interval and appearance order of the frequent patterns. Using artificial datasets, we further performed experiments to compare the extraction accuracies of the proposed and previous methods. The result shows that compared with the previous method, the proposed method allows the detection of linkage patterns correctly and comprehensively.
关键词：sequential pattern mining; linkage pattern mining; appearance interval; appearance order