A non-linear approximation to the distribution of total expenditure distribution of cruise tourists in Uruguay

This study contributes to the literature on the determinants of tourism spending on cruises at a microeconomic level, through the application of innovative methodologies framed within the machine learning literature. The objective is to study the distribution of the total expenditure of cruise passengers in Uruguay, using data of the 2016-2017 cruise season survey (collected by the Ministry of Tourism of Uruguay). Due to the nature of this variable, we implement a two stages modeling strategy. In the first stage, we model the probability of spending, and in the second, the strictly positive spending. The paper analyze the distribution of conditional expenditure to a set of sociodemographic, travel, contextual and satisfaction variables applying non-linear regression techniques with Lasso penalty and nonparametric techniques such as Random Forest. The empirical results show that the key variables that determine the average spending of cruise tourists are their residence and the port of arrival of the cruise. The analysis of the predictive performance of the models (applied through a training sample and a test sample) shows that Random Forest method has the greater predictive capacity. Finally, the importance variable is analyzed by Random Forest.