Model selection and estimation for technological growth curves

Abstract

This paper deals with the selection and estimation of the appropriate form of a growth curve for technological forecasting. Included in the model selection process are the determination of the shape of the growth curve (logistic or Gompertz) and the structure of the error underlying the model. The estimation phase entails the employment of discounted least squares to derive estimates of the model parameters that provide the best forecast. Both simulated and actual data series are employed to test the proposed algorithms for model selection and estimation. Results indicate improvement in the forecasts by employing both the selection algorithm and the discounted least squares estimation procedure.