Bus passenger flow prediction using BO-optimized XGBoost

Passenger flow prediction is essential for bus operation management. This study applies the extreme gradient boosting machine (XGBoost) to predict hourly bus passenger flow. Moreover, Bayesian optimization (BO) is adopted for parameter tuning to optimize the performance of the XGBoost. The relative importance and SHAP value are adopted to interpret the results. A comparative study with support vector machine (SVM), k-nearest neighbors (KNN), and random forest (RF) confirms that the proposed model BO-XGBoost outperforms these models in terms of accuracy and robustness. The experiment results also indicate that the BO is superior to random search.