La prévision ARMMI multivariée de chronique irrégulières = Multivariate ARIMA forecasting of irregular time series / by Pierre A. Choquette and Robert Lamy. : CS11-614/85-33-PDF

"The paper shows how smoothing filters can be built into multivariate ARIMA models. The technique can be especially useful for time series with sizable irregular fluctuations. These tend to blurr the relationships between the series, to disturb the forecasts and to underestimate the forecasting performance. The approach proposed circumvents these problems, and is illustrated by joint ARIMA forecasting of the Canadian Composite Leading Indicator and the Index of Industrial Production"--Abstract.

Lien permanent pour cette publication :
publications.gc.ca/pub?id=9.837651&sl=1

• Format MARC XML
• Format MARC HTML

Demander des formats alternatifs