基于符号表示的时间序列分类综述  

Review of time series classification based on symbolic representation

作　　者：武天鸿 翁小清[1] 单中南 WU Tian-hong;WENG Xiao-qing;SHAN Zhong-nan(College of Information Technology,Hebei University of Economics Business,Shijiazhuang Hebei 050061,China)

机构地区：[1]河北经贸大学信息技术学院

出　　处：《河北省科学院学报》2019年第3期11-20,共10页Journal of The Hebei Academy of Sciences

摘　　要：一种有效的时间序列符号表示方法不仅可以实现数据高效降维,且对噪声具有鲁棒性,对于提高分类算法的性能和效率具有重要意义。本文从基于趋势特征、基于聚类或进化计算、基于文本以及基于频率域的角度,对近年来基于符号表示的时间序列分类方法的研究进行比较归类,为研究者了解最新的符号时间序列分类方法研究动态和发展趋势提供参考。An effective time series symbolic representation method can not only achieve efficient data dimensionality reduction,but also robust to noise,which is of great significance for improving the performance and efficiency of the classification algorithm.This paper compared the research on time series classification methods based on symbolic representation in recent years from perspectives that trend characteristics-based,cluster or evolutionary calculation,text-based and frequency domain-based,and it provides a reference for researchers to understand the latest research trends and development trends of symbolic time series classification methods.

关 键 词：时间序列 符号表示方法 符号序列分类 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]