Model specifications benchmarking forecasting hotel cancellations

Abstract

This study analyzed the specifications of hotel cancellation prediction models, which is especially important at present due to the reduction in tourist flows due to intra-station pressure and global transformation processes. Studies show that the most productive are prediction models based on neural networks, as well. A comparative analysis of these models was carried out, the purpose of which is to substantiate the most effective number of canceled bookings in terms of accuracy indicators, which will optimize the forecast of the workload of the hotel business organization.