Forecasting the Global Aviation Industry Index with Machine Learning Methods

Forecasting the sectoral financial index has become increasingly important in the post-COVID-19 era, particularly in highly volatile industries such as aviation. Despite the widespread use of machine learning techniques in financial forecasting, empirical studies on their comparative performance in aviation-focused exchange-traded funds during market recovery phases remain limited. This study evaluates the predictive performance of three machine learning models: Random Forest, Facebook Prophet, and XGBoost in forecasting the U.S. Global Jets ETF (JETS). Daily closing price data from January 4, 2021, to January 10, 2025, were used. The models were evaluated using mean squared error (MSE), root mean squared error (RMSE), and mean absolute percentage error (MAPE). The empirical results show that XGBoost achieves the highest predictive accuracy, reducing RMSE by approximately 21.8% compared to Random Forest and by 64.5% compared to Facebook Prophet. While the Facebook Prophet algorithm shows weaknesses in capturing nonlinear market dynamics during structurally sensitive periods, the Random Forest method performs consistently and competitively. These results offer sector-specific evidence of the relative efficacy of machine learning techniques in aircraft market forecasting. They add to the expanding corpus of research on the use of machine learning in financial time series forecasting during periods of economically sensitive recovery.