Tourism forecasting plays a major role in tourism planning and management and it is one of the main economic activities in many countries. For this reason, it is fundamental to provide several models that allow describing and forecasting the tourist demand. International visitants who arrive at a certain tourist destination may come from countries or regions with similar or different customs and behaviours and therefore be able to present correlated arrival patterns. Based on the state-space methods with time-varying parameters, this study develops the application and comparison of univariate and multivariate models in the applied case of German and British tourist at Canary Islands (Spain). The choice of model can be conditioned by the volume of tourists from one country with respect to the other. Structural models will be used incorporating intervention and exogenous variables, among which airline seat reservations for regular flights.
A kernel density-based particle filter for state and time-varying parameter estimation in nonlinear state-space models
