Abstract
Booking hotels online is now a very common way for people to travel and stay, but a large number of cancellations due to itinerary changes and other factors can have a big impact on hotels, such as losing customers who really need a certain room type and losing them to other hotels. In order to reduce hotel losses, this paper uses the data of two hotels through data published on Kaggle's official website, identifies the factors that have the greatest impact on hotel cancellations through EDA visualization, and gives improvement measures. Machine learning algorithms are then used to guess whether the customer will cancel the booking. Each algorithm has its own area of expertise, so this article makes a comparison to the performance of decision trees, logistic regression, random forests. The result is that random forests have the highest accuracy and hotel managers can use the model to predict and change business strategies to increase profits.
Publisher
Darcy & Roy Press Co. Ltd.
Reference10 articles.
1. Kimes S E, Wirtz J. Has revenue management become acceptable? Findings from an Interna-tional study on the perceived fairness of rate fences. Journal of Service Research, 2003, 6 (2): 125 – 135.
2. Hadden J, Tiwari A, Roy R, et al. Computer assisted customer churn management: State-of-the-art and future trends. Computers & Operations Research, 2007, 34 (10): 2902 - 2917.
3. Bai Ruirui. Research on customer prediction and retention countermeasures of hotel reservation platform. Zhengzhou University, 2021.
4. Talluri K T, Van R G. The theory and practice of revenue management. New York, NY: Springer, 2005.
5. Ahlam A, et al. Cancellation Prediction for Flight Data Using Machine Learning. 2nd International Conference on Advances in Science & Technology (ICAST), 2019.