引用本文:K. D. C. Stoodley,卢润德.非连续含有准周期分量时间实例系列的实时监测[J].控制理论与应用,1986,3(1):35~44.[点击复制]
K. D. C. Stoodley ,Lu Runde.The Real-Time Monitoring of Time Series with Discontinuities And Quasi-Periodic Components[J].Control Theory and Technology,1986,3(1):35~44.[点击复制]
非连续含有准周期分量时间实例系列的实时监测
The Real-Time Monitoring of Time Series with Discontinuities And Quasi-Periodic Components
摘要点击 1104  全文点击 550  投稿时间:1984-04-28  修订日期:1985-01-18
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  
  1986,3(1):35-44
中文关键词  
英文关键词  
基金项目  
作者单位
K. D. C. Stoodley,卢润德 英国不来福大学 
中文摘要
      本文提出了一种新的线性增长型数字模型选值标准。该标准适合于分析含有准周期分量的时间序列。根据这种新标准选值,尽管时间序列含有随机的倾向变化和准周期分量,线性增长型模型仍可预报该随机过程的期望和倾向。基于双模刑法,本文还提出了一种新的披露微小倾向变化的方法。应用该办法及对阶跃变化和瞬间变化的探测,使我们有可能对含有伪周期分量的非连续时间序列进行实时的检测和预报。
英文摘要
      A new criterion is described for choosing the parameters of the forecasting algorithm corresponding to the linear growth model, where quasi -periodic components exist in the series. The forecasts based on the resulting algorithm allow prediction of the mean Level and slow trend of a process in spite of large random variations, stochastic trend changes and possible quasi -periodic components in the series. A new slope change detection method is also discussed based on a dual model approach. This method in conjunction with previously developed methods for detection of step changes and transients allows monitoring of a time series with quasi -periodic components and subject to discontinuities.