作者: Helmut Lütkepohl
DOI: 10.1016/0304-4076(84)90023-X
关键词: Arma process 、 Applied mathematics 、 Process (engineering) 、 Mathematics 、 Linear map 、 Stochastic process 、 Mathematical optimization 、 Univariate
摘要: Abstract Linear transformations of stochastic processes are used in many ways economic analyses, for example when linear aggregates or subprocesses considered. It is demonstrated that a transformation vector ARMA process again an and conditions stationarity given. Three different predictors linearly transformed compared. Forecasting the original transforming predictions superior to forecasting directly univariate components process. Conditions equality three forecasts provided.