Predictive models for hotel booking cancellation: a semi-automated analysis of the literature

Abstract

In reservation-based industries, an accurate booking cancellation forecast is of foremost importance to estimate demand. By combining data science tools and capabilities with human judgement and interpretation, this paper aims to demonstrate how the semiautomatic analysis of the literature can contribute to synthesizing research findings and identify research topics about booking cancellation forecasting. Furthermore, this works aims, by detailing the full experimental procedure of the analysis, to encourage other authors to conduct automated literature analysis as a means to understand current research in their working fields. The data used was obtained through a keyword search in Scopus and Web of Science databases. The methodology presented not only diminishes human bias, but also enhances the fact that data visualisation and text mining techniques facilitate abstraction, expedite analysis, and contribute to the improvement of reviews. Results show that despite the importance of bookings’ cancellation forecast in terms of understanding net demand, improving cancellation, and overbooking policies, further research on the subject is still needed.