Libro

Descripción

Shipping is under increasing pressure to improve its environmental performance, but most emissions reporting remains retrospective. This book explores whether historical vessel-level emissions records can be transformed into a forward-looking early-warning system for operational carbon intensity deterioration. Using verified European Union Monitoring, Reporting and Verification (EU MRV) data from 2018 to 2024, the study develops a longitudinal dataset covering thousands of vessels and evaluates multiple statistical and machine-learning approaches. Logistic Regression, Random Forest and XGBoost models are compared using strict time-ordered validation, with the final year preserved as an out-of-time holdout. The analysis demonstrates the value of vessel history for identifying ships at greater risk of next-year deterioration and introduces a practical risk-ranking framework for fleet-level screening. Particular attention is given to model explainability, calibration, robustness and reproducibility. The book is intended for researchers, maritime professionals, shipowners, regulators and practitioners interested in maritime decarbonization, emissions analytics and data-driven environmental risk management.