Abstract:

A management information system, or MIS, is a computer system that provides managers with the tools to plan, organize, manage, and control business processes. Various input data can be implemented in MIS. From the point of view of hotels and other tourism companies, knowledge of the demand for tourism is important. The point is that predicting the number of tourists is important for destination management and marketing. The demand for tourism is subject to many factors, such as seasonality. Destination managers must be able to anticipate demand in their businesses so that they can plan certain processes, such as determining the number of employees on duty on a specific day or even a specific hour. While traditional time series forecasts may take into account factors such as seasonality. We have large amount of sources of big data to increase the accuracy of forecasts. The aim of this paper is to compare the performance of the Bayesian and classical approaches to time series modelling in order to find a suitable model for forecasting tourism demand. The results show that BVAR models outperform the predictive performance of VAR model. The BVAR model with different priori densities achieve similar performances in predicting tourism demand. The BVAR model can therefore be applied to MIS, where the outputs of these models can serve as input data in modelling and analysis of decision-making processes.