Accuracy of Comparison of Tourism Demand Forecasts in Olympic Hosts

Abstract

Precise forecasts concerning tourism demand for a certain place serve as support for correct decision making by managers, especially when the realization of a major event directly affecting urban and economic local development is considered. The aim of this study understands the behavior of tourism demand regarding the holding of one of the biggest mega-events of our days. In this regard, the tourism demand of two countries that hosted the Olympic Games was modeled and forecasted using two different methodologies, a linear and a non-linear and the resulting forecasts were combined utilizing five methods – three linear and two non-linear. In both series, the individual forecasts obtained by the ANN model were more accurate than the ones generated by the SARIMA model and the combination of these forecasts, especially considering the linear regression method, further increased the precision of the aforementioned combinations. No impact in demand was observed because of the holding of the Olympic Games since the event took place during the high season, when the chosen molds displayed more exactitude.