基于统计特征的时序数据符号化算法

为克服SAX(符号聚合近似)算法对时序信息描述不完整的缺陷,提出基于统计特征的时序数据符号化算法,与SAX不同的是,该算法将时序符号看作矢量,而各时序子段的均值和方差则分别作为描述其平均值及发散程度的分量.由于该算法能够比SAX提供更多的描述信息,因而在时序数据挖掘应用中能够获得比SAX更精确的结果.大量的实验也证实了它的出色表现.

The Symbolic Algorithm for Time Series Data Based on Statistic Feature

A new symbolic algorithm for Time Series Data Based on Statistic Feature is put forward in order to surmount the bugs with which SAX(Symbolic Aggregate Approximation) Algorithm can not describe time series information fully.This algorithm,differing from the SAX,considered the symbolic as vector,and Mean and variance from each subsequence were regarded as components by which its mean value and radiation degree are described respectively.Since it could provide more information described time series than SAX do,more accuracy result could be get when it is applied to time series data-mining.Its excellent behave.have been proved by a lot of experiments.