论文信息 - Model Selection for Multi-Step-Ahead Forecasting

Model Selection for Multi-Step-Ahead Forecasting

Abstract It is a common practice to fit a single time series model for the purpose of forecasting an observed time series at several different leads. By examining autoregressivc models, we demonstrate both theoretically and empirically that there are situations in which better forecasts can be obtained by using a different “model” for each lead.

D. Findley | D. F. Findley

[1] R. Bhansali. Effects of Not Knowing the Order of an Autoregressive Process on the Mean Squared Error of Prediction—I , 1981 .

[2] Ruey S. Tsay,et al. Order Selection in Nonstationary Autoregressive Models , 1984 .

[3] W. Fuller,et al. Properties of Predictors for Autoregressive Time Series , 1981 .

[4] H. L. Gray,et al. New arma models for Wolfer’s sunspot data , 1978 .

[5] Genshiro Kitagawa,et al. An approach to the prediction of time series with trends and seasonalities , 1982, 1982 21st IEEE Conference on Decision and Control.

[6] R. Shibata. Asymptotically Efficient Selection of the Order of the Model for Estimating Parameters of a Linear Process , 1980 .

[7] Paul Kabaila,et al. Estimation based on one step ahead prediction versus estimation based on multi-step ahead prediction , 1981 .