Predicting Hotel Booking Cancellations using Machine Learning Techniques

This study introduces a comprehensive machine learning system to enhance the accuracy of hotel reservation cancellation forecasts. By leveraging a diverse dataset encompassing factors like booking timelines, customer demographics, room preferences, and past cancellation rates, we employ advanced machine learning techniques such as Logistic Regression (LR), Decision Tree Classifier (DTC), Random Forest Classifier (RFC), Gradient Boosting Classifier (GBC), Support Vector Machine (SVM), and Artificial Neural Network (ANN) to craft predictive models. Rigorous evaluation using metrics like recall, F1-score, accuracy, and precision demonstrated the effectiveness of these models. The ANN achieved the highest performance with an accuracy of 99.02%, a recall of 97%, a precision of 99%, and an F1-Score of 99%. DTC and RFC also showed strong results, while LR and SVM had relatively lower performance. Integrating these models into a cohesive system allows for real-time prediction and management of booking cancellations, providing hotel management with a valuable tool for optimizing operations, enhancing revenue forecasting, and improving customer satisfaction. This study highlights the potential of machine learning to transform hotel booking management through data-driven decision-making.