Improving Flight Ticket Prediction Performance On Random Forest Machine Learning Model Using Bayesian Optimization

The rapid development of the aviation industry has increased the need for accurate airline ticket price prediction, benefiting both airlines and consumers. This study evaluates the performance of the Random Forest model before and after Bayesian Optimization to enhance predictive accuracy. Using a dataset from "Ease My Trip," the model is assessed with Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-squared Score (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>). Hyperparameter Tuning is performed using Bayesian Optimization to optimize model parameters. Results indicate that tuning improves the model’s ability to capture data patterns, enhancing predictive performance. The untuned model achieved an MSE of 4,910,040.02, RMSE of 2,215.86, MAE of 829.00, and an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> Score of 0.9903918. After tuning, optimal parameters (max_depth: None, min_samples_leaf: 1, min_samples_split: 5, n_estimators: 100) improved MSE to 4,815,187.66, RMSE to 2,194.35, and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> Score to 0.9905774, though MAE slightly declined to 850.42. This study highlights the effectiveness of Hyperparameter Tuning in refining model performance, offering valuable insights for airline pricing strategies and consumer decision-making. The findings suggest that optimized models can enhance revenue management for airlines and improve fare forecasting for passengers. Future research could explore alternative methods for more efficient tuning, such as Early Stopping method with Bayesian Optimization.