【摘要】 本文引入平稳时间序列分类的最近邻准则法.新时间序列用最近邻准则分类.利用Kullback-Leibler信息量计算时间序列间的相异测度.本文的新结果是:给出两平稳时间序列间K-L信息量的定义,导出MA(q)及ARMA(p,q)时间序列的相应公式.相应公式推广到多维情况.本文结果在机械工艺实践中获得应用.更多还原

【Abstract】 A nearest neighbor rule approach is introduced for the classi’ication of Stationary time Series. The new time series is cla^sfied by nearest neighbor rules. The dissiralarity measure between the time series is Computed by the use of the Kullback-Leibev information number. New results in this paner are the definition of K-L information number between two stationary time Series and corresponding formulae of MA.(q) and ARMA(p,q) time series. The Corresponding formulae are extended to the multivariate case. The results in this paper have been applied to the practice of machinary technology.更多还原