Forecasting Tourist Arrivals: Nonlinear Sine Wave or ARIMA?

Abstract

This article builds a model for data series that have both seasonal and nonseasonal ARIMA patterns. In this type of model, the seasonal and nonseasonal elements are multiplied by each other. The model is then used to fit the tourist arrival data. Specifically, this approach is applied to Singapore tourist arrival data, and the fitted model is then used to forecast the tourist arrivals. It is shown that the proposed model generates forecasts with lower error magnitudes than a sine wave nonlinear time-series model. Policy implications are also addressed.