Resumen de Univariate Periodic and Nonperiodic Modeling of Tourism Time Series Compared

Lindsay W Turner, Austin M. M. Kijagulu

• Seasonality is a critical issue in tourism forecasting. The method of dividing the data into four seasonal series prior to forecasting is examined by comparison with the alternative nonperiodic series for total travel into Japan from all source regions for quarterly data between March 1978 and December 1995. This in-depth analysis attempts to provide a conclusive test of whether the periodic approach is or is not appropriate as a method for analyzing international tourism series. The results indicate conclusively that periodic modeling does not improve forecasting accuracy, and suggest that integrative models and in particular the structural model perform best as tourism time series forecasting methods.