Forecasting with a dynamic regression model: a heuristic approach
1982
Extract: A procedure is presented to facilitate the use of a bivariate time series model for forecasting hog prices. Detection of causality from the innovations of an ARIMA-filtered sows farrowing series to the innovations of an ARIMA-filtered hog price series suggested improved forecasts should be possible with a dynamic regression model. Forecasts from the univariate and bivariate models were generated for 20 quarters. A statistical procedure for evaluating the difference in the out-of-sample mean squared errors was applied. The improvement in forecasting accuracy from the bivariate model was statistically low.
Mostrar más [+] Menos [-]Palabras clave de AGROVOC
Información bibliográfica
Este registro bibliográfico ha sido proporcionado por National Agricultural Library