Predicting Vessel Speed Over Ground: A Machine Learning Approach for Enhancing Maritime Transport

As global maritime transport evolves, building resilient transport systems necessitates the use of advanced predictive technologies. This research develops machine learning models for vessel speed over ground prediction, addressing critical challenges in maritime transport reliability. Using comprehensive Automatic Identification System data, the study examines neural networks, tree-based models, and Transformer architectures to assess their predictive capabilities. Focusing on the St. Lawrence River, a strategically significant maritime corridor, the research demonstrates how precise speed prediction enhances operational risk mitigation, logistical planning, and systemic adaptability. The proposed approach offers a scalable solution for developing more responsive maritime transport networks.