Several studies have provided a review of past tourism demand forecasting
and a comprehensive analysis of forecasting model selection in terms of error
magnitude and directional change error (Witt and Witt, 1995; Witt, Song, and
Louvieris, 2003). Previous tourism demand studies examined various forecasting
models for both inbound and outbound tourism for different destinations in different
time periods.
Martin and Witt (1989) and Sheldon (1993) compared the accuracy of seven
forecasting methods for simulating visitor flows among twenty-four origin-destination
pairs. They found that exponential smoothing was the second most accurate model in
terms of mean absolute percentage error (MAPE). Also, advance forecasting models
have been widely used for tourism demand forecasting in the past four decades with