时间序列的表示与分类算法综述

Review of Time Series Representation and Classification Techniques

时间序列是按照时间排序的一组随机变量,它通常是在相等间隔的时间段内,依照给定的采样率,对某种潜在过程进行观测的结果。时间序列数据广泛地存在于商业、农业、气象、生物科学以及生态学等诸多领域,从时间序列中发现有用的知识已成为数据挖掘领域的研究热点之一。在时间序列表示方面,主要介绍了非数据适应性表示方法、数据适应性表示方法和基于模型的表示方法;针对时间序列的分类方法,着重介绍了基于时域相似性、形状相似性和变化相似性的分类算法,并对未来的研究方向进行了进一步的展望。

Time series is a set of random variables ordered in timestamp.It is often the observation of an underlying process,in which values are collected from uniformly spaced time instants,according to a given sampling rate.Since time series data exist widely in various application domains,such as finance,agriculture,meteorology,biological science,ecology and so on,discovering knowledge from time series has become one of the mainly research fields of data mining.In this paper,a comprehensive review on the existing time series representation and classification research was given.In the term of time series representation,three different categories named non-data adaptive,data adaptive and model based were summarized.A summary of several time series classification method,namely similarity in time,similarity in shape and similarity in change was also provided.