Forecasting tourism: a combined approach

Abstract In this article, we employ a combined seasonal nonseasonal ARIMA and sine wave nonlinear regression forecast model to predict international tourism arrivals, as represented by the number of world-wide visitors to Singapore. Compared with a similar study of the accuracy of international tourist arrivals forecasts by Chan (Journal of Travel Research, 1993, 31, 58–60)1 and Chu (Journal of Travel Research, 1998, 36, 79–84)2 using other univariate time series models, our proposed model has the smallest mean absolute percentage error.