| Summary: | Booking cancellation prediction becomes more significant than before, which impacts
decision making in the hospitality industry. In the revenue management system, with
inaccurate prediction of hotel demand, overbooking and cancellation policy might lead to a
negative influence on the operation of the hotel and reputation of the hotel.
Using PMS data and external data, addressing the booking cancellation problem as a
classification problem, the author used different algorithms and different models to get the
highest accuracy, the result was exceeding 0.89, which shows the hotel industry can predict
whether a booking is likely to be canceled with high accuracy.
Models allow the hotel industry to make different actions on overbooking and cancel booking
based on which factors were most important in the model. A high accuracy model can prevent
the enterprise from reputation and profit losing. Moreover, there will be some future research
suggestions provided in the end.
Keyword: Data science, booking cancellation, hospitality industry, machine learning,
predictive modeling, revenue management, weather, distance, classification
|
|---|