PREDICTIVE MODELING OF BOOKING CANCELLATIONS IN HOSPITALITY USING EXPLAINABLE AI

Abstract

The hospitality industry faces significant challenges due to booking cancellations, which directly impact revenue management, resource allocation, and operational efficiency. With the rise of online booking platforms and flexible cancellation policies, the unpredictability of customer behavior has increased, making accurate forecasting of cancellations a critical requirement. Traditional statistical methods often fail to capture complex patterns in booking data, limiting their predictive capabilities. This research proposes an interpretable machine learning framework for forecasting booking cancellations in the hospitality sector, focusing on both prediction accuracy and model transparency. The proposed approach integrates advanced machine learning techniques such as decision trees, gradient boosting, and explainable artificial intelligence (XAI) methods to provide insights into the factors influencing cancellations. Unlike black-box models, the framework emphasizes interpretability through feature importance analysis, SHAP (SHapley Additive exPlanations), and rule-based explanations, enabling hotel managers to understand and trust the model’s predictions. The system utilizes historical booking data, including customer demographics, booking lead time, pricing, seasonality, and market segment, to identify patterns associated with cancellations. Additionally, data preprocessing techniques such as feature engineering, handling missing values, and normalization are applied to improve model performance. Experimental results demonstrate that the proposed model achieves high accuracy and improved interpretability compared to traditional approaches, enabling better decision-making in revenue management and operational planning. The framework also supports real-time prediction, allowing hotels to take proactive measures such as overbooking strategies, targeted promotions, and personalized customer engagement. By combining predictive analytics with interpretability, the proposed system bridges the gap between advanced machine learning and practical decision-making in the hospitality industry. This research highlights the importance of explainable models in enhancing trust, transparency, and efficiency in hotel management systems, ultimately contributing to improved customer satisfaction and business performance