Freight Mode Choice

Freight mode choice modeling aims to predict which transportation mode (e.g., truck, rail, ship) will be selected for a given shipment, considering factors like distance, commodity type, and shipper characteristics. Recent research heavily utilizes machine learning, particularly ensemble methods like Random Forests and boosting algorithms, achieving high prediction accuracy by incorporating spatial data and employing techniques like local modeling for specific commodity types. These advancements improve the efficiency and accuracy of freight logistics, with practical applications demonstrated by significant cost savings in commercial settings through improved rate forecasting, for example, using novel methods like signature transforms. The field's focus is on enhancing model accuracy, interpretability, and the incorporation of diverse data sources to better understand and optimize freight transportation networks.