Hotel Booking Analysis: Customer Segmentation and Demand Forecasting

Abstract

This study presents a thorough analytical approach to hotel booking data. By using machine learning and time series methodologies, the research integrates customer segmentation, cancellation prediction, and demand forecasting. The study used K-Means, HDBSCAN clustering to define guest segments, Random Forests with SMOTE to classify booking cancellations, and SARIMA models to forecast future demand. The findings demonstrate that this approach leads to more accurate demand predictions and a deeper understanding of booking patterns. This improved understanding can help hotels optimize their resource allocation, pricing strategies, and customer targeting.