Authors : M.A. Umar
Abstract: In this study, we examined the Holt-winter, double exponential and the linear regression trend parameter estimation techniques and compare their forecast quality via the criteria of forecast as in Gilchrist, using time’s series data of exchange rates of the Naira to the dollar. Forecasts using these methods are presented and comparison statistics and statistics of errors for the methods are examined. It is found that the Holt-winter forecasting method with choice of smoothening constant a = 0.2 and b = 0.5 produced better forecasts than the rest of the methods. In general, the Holt-winter always gives outstanding forecasts than the other methods, as seen on the table of comparison statistics and statistics of error. Also, the required constraint on the smoothening constants for the Holt-winter to meet the criteria of forecast is suggested.
M.A. Umar , 2007. Comparative Study of Holt-Winter, Double Exponential and the Linear Trend Regression Models, With Application to Exchange Rates of the Naira to the Dollar . Research Journal of Applied Sciences, 2: 633-637.