时间序列数据挖掘的相似性度量综述

发布时间：2018-05-09 04:24

本文选题：时间序列数据挖掘 + 时间序列相似性　；参考：《控制与决策》2017年01期

【摘要】：在时间序列数据挖掘中,时间序列相似性是一个重要的概念.对于诸多算法而言,能否与一种合适的相似性度量方法结合应用,对其挖掘性能有着关键影响.然而,至今仍没有统一的度量相似性的方法.对此,首先综述了常用的相似性度量方法,分析了各自的优点与不足;其次,讨论了近年来出现的时序相似性的新解释及其度量方法;再次,探讨了相似性度量在时序挖掘任务中的应用以及与挖掘精度的关系;最后给出了关于时序相似性度量进一步的研究方向.
[Abstract]:Time series similarity is an important concept in time series data mining. For many algorithms, whether they can be applied in combination with an appropriate similarity measurement method has a key impact on their mining performance. However, there is still no uniform method to measure similarity. In this paper, first of all, the common similarity measurement methods are summarized, and their advantages and disadvantages are analyzed. Secondly, the new explanation of temporal similarity and its measurement methods are discussed in recent years. Thirdly, This paper discusses the application of similarity measurement in time series mining task and the relationship between similarity measurement and mining precision, and finally gives the further research direction of time series similarity measurement.
【作者单位】：南京航空航天大学计算机科学与技术学院;南京航空航天大学软件新技术与产业化协同创新中心;
【基金】：国家自然科学基金项目(61501229) 中央高校基本科研业务费专项资金项目(NS2015091,NJ20160013)
【分类号】：O211.61;TP311.13

【相似文献】