Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{209084,
        author = {Nurudeen A. Raji and Medekannu M. Ogunbiyi and Abdullateef O. Bakare and Choice Eseoghene and Ojo Seun},
        title = {Machine Learning Analysis for Fuel Consumption in Heavy Vehicles},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {5},
        pages = {678-687},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=209084},
        abstract = {Efficiently managing heavy vehicle fuel consumption is still a top priority for the transportation sector. With the use of cutting-edge machine learning techniques, this research aims to develop a robust model that can predict fuel consumption with improved accuracy given a variety of features from different heavy vehicles. The project uses a data-driven approach and draws on a sizable real-world dataset from a variety of heavy vehicle. To assure data integrity and coherence, this dataset has undergone painstaking processing and improvement. We explore the dataset's intricate patterns and using exploratory data mining techniques. To identify the most relevant predictors impacting fuel consumption dynamics, feature selection and engineering are applied as integral parts of the process. A wide range of machine learning techniques such as regression analysis, ensemble learning (voting and stacking regressor), hyperparameter tuning, neural network dropout and batch normalization underwent careful training and evaluation. With thorough evaluation, we identify stacking ensemble model as the front-runner with a stunning accuracy of 0.999983 and MSE (mean square errors) of 0.001007 and (support vector machine) SVM,  XGBoost, gradient boosting, and random forest by a rigorous comparative analysis, revealing them to be the best performing among the models after hyperparameter tuning 0.999968% , 0.999525%, 0.99821%,  0.996388% 2, and MSE of 0.00186, 0.028248, 0.105822,  0.214788 respectively. Our analysis underscores the importance of testing multiple techniques before drawing definitive conclusions. Particularly, the Support Vector Machine (SVM) initially exhibiting poor performance with an accuracy of 23.36% and MSE of 45.57, highlighting the necessity for thorough evaluation.},
        keywords = {fuel consumption prediction, machine learning, heavy vehicles, regression analysis, ensemble Learning.},
        month = {October},
        }