Fuel Efficiency Prediction Using Machine Learning, A TensorFlow-Based Approach to Enhancing Automotive Sustainability

Fuel efficiency prediction model with TensorFlow, analyzing key vehicle parameters to better inform sustainable auto design and consumer decisionmaking. This model uses the dataset of auto mpg and includes preprocessing of the data and the building of a deep learning model in TensorFlow’s Sequential API as the basis for evaluating metrics including Mean Absolute Error. Existing linear models often miss complex relationships and struggle with large datasets. In contrast, TensorFlow deals with non-linear interactions. The model yields a training accuracy of about $80-90 \%$ and thus performs well. However, the validation accuracy is about $60-70 \%$. The most important factors affecting fuel efficiency are found to be vehicle weight, engine displacement, and number of cylinders. This research suggests the possibility of improving better fuel efficiency prediction and thereby helping manufacturers, policymakers, and consumers to conserve energy and design better automobiles with the help of machine learning technologies, as well as serving as a model for future energy-saving automobile technologies.