aceAn integrated method of advanced optimisation and adaptive ensemble learning for ship fuel consumption prediction
• Proposes an adaptive cluster-based multi-ensemble (ACME) model for robust ship fuel consumption prediction. • Develops a SHAP-weighted multi-model feature selection (SWFS) algorithm for dimensionality reduction. • Introduces a hierarchical adaptive parameter space exploration (HAPSE) method for efficient tuning. • Establishes a dual-layer SHAP framework for global and local model interpretability. • Integrates multi-source data fusion to enhance prediction accuracy and operational relevance. Accurate predictio…
"Oiled Bird - Black Sea Oil Spill 11/12/07" by marinephotobank is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/.
By April 2026, researchers had developed and tested an integrated framework combining advanced optimisation with adaptive ensemble learning for ship fuel consumption prediction. The system fused noon reports, AIS, and meteorological and oceanographic reanalysis data, applied SHAP-weighted feature selection and hierarchical parameter search, and used cluster-based multi-ensemble learning to adapt to different operational conditions.
The work matters because accurate, interpretable fuel consumption prediction directly supports fuel cost reduction and decarbonisation efforts in shipping, a hard-to-abate sector. What remains uncertain from the supplied text is how the model performs prospectively in live fleet deployment, across diverse vessel types and routes, and whether efficiency gains translate into measured emissions reductions outside the experimental evaluation.
- Multi-source data fusion used spatio-temporal alignment to integrate ship noon reports, Automatic Identification System data, ECMWF Reanalysis v5, and Global Ocean Physics Analysis and Forecast data
- SHAP-based Weighted Feature Selection algorithm combined multi-model SHapley Additive exPlanations value assessment with recursive feature elimination to remove redundant features
- Hierarchical Adaptive Parameter Space Exploration combined global random search and local grid search for hyperparameter optimisation
- Adaptive Cluster-based Multi-Ensemble model used data clustering and model fusion to capture operational heterogeneity and adaptively assign weights across clusters
The integrated ACME framework increased ship fuel consumption prediction accuracy and generalisation, outperforming mainstream models and providing data-driven tools for energy efficiency management and decarbonisation in maritime transport.
The rundown
The pipeline was structured in four technical components: multi-source data fusion for feature space construction, SHAP-weighted feature selection for dimensionality reduction, hierarchical adaptive parameter space exploration for tuning, and adaptive cluster-based multi-ensemble for handling operational heterogeneity.
Evaluation reported in the April 2026 peer-reviewed article found the proposed model outperformed six mainstream machine learning models and three classical ensemble methods across multiple metrics, with SHAP-based interpretability used to quantify feature contributions and reveal effects of value changes.
Source code was made publicly available at the linked GitHub repository, positioning the framework as a tool for sustainable maritime operations rather than a purely theoretical advance.
Sources
- Peer-reviewedTransportation Research Part C: Emerging Technologies2026-04-02
How should this claim be treated?
The debate