Identifying critical hotel cancellations using artificial intelligence

Abstract

Cancellations have a significant impact on the hotel and lodging industry because they directly affect income and are thus considered critical in revenue management. Specifically, cancellations made close to the time of service are the most damaging for hotels because they leave management with no time to react. The use of Personal Name Records (PNR) has led to new approaches in this field, however despite this novel research area there are no investigations focusing on forecasting for individual hotel cancellations made close to the time of service. With the aim of filling this gap, this research is intended to identify those individuals likely to make cancellations in a short-horizon of time using Artificial Intelligence (AI) techniques through PNR data. Promising results have been achieved with 80% accuracy for cancellations made 7 days in advance. By taking this approach, booking management systems, as well as cancellation policies may be optimised.