Abstract
Tourism, one of the biggest industries in many countries, has been considered a complexly integrated and self-contained economic activity.As key determinants of thetourism demand are not fully identified to some extent different forecasting models vary in thelevel of accuracy. By comparing the performance of diverse forecasting models,including the linear regression, autoregressive integrated moving average (ARIMA), Grey model, and their joint Fourier modified models, this paper aims at obtaining an efficient model to forecast the tourism demand. As a result, the accuracy of the conventional models is found to be significantly boosted with the Fourier modification joined. In the empirical case study of international tourism demand in Taiwan, the Fourier modified seasonal ARIMA model FSARIMA (2,1,2)(1,1,1)12 is strongly suggested due to its satisfactorily high forecasting power. We further employ the model to provide the Taiwan’s tourism demand in 2013 so as to assist policy-makers and related organizations in establishing their appropriate strategies for sustainable growth of the industry.