Amid rising environmental concerns and escalating fuel costs, enhancing vehicle fuel efficiency is a critical focus in automotive engineering. This study applies advanced ML (machine learning) approaches to predict vehicle fuel efficiency, specifically targeting the ‘comb08’ variable in a comprehensive vehicle dataset. Six predictive models are evaluated: Linear Regression, Decision Trees, and Random Forest among them. Rigorous data preprocessing ensures data quality and consistency, involving normalizing features, handling missing values, and encoding categorical variables. After preprocessing, models are trained and validated to assess their predictive accuracy and robustness. The performance varies significantly across models, with the Decision Tree model standing out as the most accurate and robust, achieving a low root mean square error (RMSE) of “ 2.752455 “ in fuel efficiency predictions. These findings enhance our understanding of the factors influencing fuel efficiency and provide essential insights for developing more energy-efficient vehicles. The implications extend beyond academia, impacting automotive design and informing environmental policies. By showcasing the potential of machine learning, this research underscores its pivotal role in advancing fuel economy standards and promoting sustainability within the automotive industry. It highlights the importance of leveraging advanced analytical techniques to address critical challenges in modern transportation systems, contributing to more sustainable and cost-effective automotive solutions.
Leveraging Machine Learning for Predicting Vehicle Fuel Efficiency
06.09.2024
1775472 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Predicting Distillation Properties of Gasoline Fuel Blends using Machine Learning
British Library Conference Proceedings | 2022
|Hybrid vehicle fuel efficiency using inverse reinforcement learning
Europäisches Patentamt | 2015
|