Hotel cancelation predictions

Hotel Booking Intelligence – Predictive Cancellation Modeling

Developed a Machine Learning project to analyze hotel booking behavior and predict reservation cancellations. The project includes data cleaning, exploratory data analysis, feature preparation, and benchmarking of five classification algorithms: Logistic Regression, KNN, Decision Tree, Random Forest, and XGBoost.

XGBoost achieved the strongest performance in the presented benchmark with 84.2% Accuracy, 82.5% Precision, 81.1% Recall, and 81.8% F1-Score. The model can be used as an early risk signal to identify reservations with a higher likelihood of cancellation and support data-driven revenue management decisions.