A COMPARISON OF NONLINEAR TIME SERIES STATISTICAL FORECASTING TECHNICS

Abstract

Objective of this research was to comparison efficiency of statistical forecasting technics between Box-Jenkins method and Radial Basis Function method for four model nonlinear time series viz model : Quardratic , Cubic , Exponential and Unpattern. The substitute data for four nonlinear time series models for this study used are the sales volume of compresser and motorcycle in Thailand, the import volume of mineral fuel and lubricant, and the export volume of pigs in Thailand. The results of this study show that Radial Basis Function model has a suitable model and more accuracy than Box-Jenkins Method in case of model nonlinear time series