Modelling, forecasting and testing decisions for seasonal time series in tourism

Abstract

Time series analysis for basic tourism parameters, in the countries of the Balkan Peninsula, have been emphasized in recent research. Moreover, some of them have also shown a trend, aside from the rising variance during the period-heteroscedasticity. All of these characteristics of a time series of tourist demand, result in them being a great challenge for modeling. Therefore, there are different types of models that can be implemented for the modeling of a time series, which include accentuated seasonal components. Throughout this paper, multiple tests are performed using several parameters of the time series, with the ARIMA model, in an attempt to find any influence on the fit and validity of the model. For the accepted models, series are predicted for a year in advance and, in addition, a method of testing the decisions made by authorities in the field of tourism is presented.