Abstract: In this paper, a versatile nonparametric nonlinear time-series model, viz. Functional-coefficient autoregressive (FCAR) model, in which the coefficient function changes gradually rather than abruptly, is considered. As an illustration, this model is applied for modelling and forecasting of India's annual export lac data during the period 1900 to 2000. Comparison of the performance of FCAR model vis-à-vis the Self exciting threshold autoregressive (SETAR) and Autoregressive integrated moving average (ARIMA) models is also made from the viewpoint of dynamic one-step and two-step ahead forecasts along with Mean square prediction error (MSPE), Mean absolute prediction error (MAPE) and Relative mean absolute prediction error (RMAPE). The SAS, Ver. 9.1 and SPSS software packages are used for data analysis. Superiority of FCAR model over SETAR and ARIMA models is demonstrated for the data under consideration.
Keywords: Akaike information criterion, annual lac export data, ARIMA, FCAR and SETAR models