Enhancing Airline Passenger Demand Forecasting in Australia: Machine Learning Models with Bayesian Optimization

Accurate forecasting of airline passenger demand is vital for strategic planning and operational efficiency in the aviation industry. While traditional models have been foundational, emerging machine learning (ML) techniques and hyperparameter optimization methods offer significant improvements in predictive performance. We evaluate advanced ML models—Support Vector Machines (SVM), Artificial Neural Networks (ANN), Generalized Additive Models (GAM), Gaussian Process Regression (GPR), and Regression Trees—integrated with Bayesian optimization for hyperparameter tuning. Results demonstrate that Bayesian optimization substantially improves model efficiency and precision, with GAM exhibiting superior performance due to its flexibility in capturing nonlinear trends without extensive tuning. The study underscores the importance of integrating macroeconomic variables and structured modeling approaches for robust predictions.