In the leisure, tourism and mobility markets, customer volumes fluctuate short-term. It is difficult for SMEs in this sector to estimate customer volumes with sufficient precision and to react appropriately to fluctuations. The result can be a misallocation of resources. As part of the Innosuisse project "Short-term forecasting and control of the number of guests in the leisure and tourism market", econometric models and methods of machine learning were evaluated in order to predict the number of visitors for SMEs in the leisure, tourism and mobility sectors. The project has shown that the data of a single company are not enough to predict the fluctuation with sufficient accuracy.
In this project, we examine the hypothesis that a more informative dataset can be created through the merger of several company data in the leisure/tourism and mobility sector due to the correlation effects. We want to show that this new database can significantly increase the quality of forecasting for the individual company.
In addition to linear econometric models, we use machine learning methods such as support vector machines and neural networks to predict the customer volume on a daily/ hourly basis.
Historical data from nine companies in the Lucerne area for the period 2007 - 2018 are available for this project.