An Analysis for Car Fuel Estimation with Regression Methods

Abstract

Fuel consumption and efficiency have emerged as pressing concerns in the context of growing energy sources and increasing environmental awareness. Machine learning, a subset of artificial intelligence, leverages intricate data structures and variable information to make predictions. These algorithms play a pivotal role in modeling and forecasting across diverse industries like healthcare, finance, banking, and energy. This study offers a comprehensive overview of a typical machine learning project flow, with a particular focus on fuel prediction. The project encompasses key stages such as data collection, data preparation, model development, and evaluation. The methodologies and algorithms employed in this research hold the potential for broader applications in various forecasting projects and industry sectors. In this investigation, fuel estimation was carried out using a set of features from the Auto MPG Data Set, sourced from the University of California. These features included Mpg (fuel consumption), Number of Cylinders, Engine Volume, Horsepower, Vehicle Weight, Acceleration, Model Year, Vehicle Origin, and Vehicle Name. Various regression algorithms, namely Linear Regression, Ridge Regression, Lasso Regression, and XGBoost, were applied to predict fuel consumption. The study's outcomes were generated by splitting the dataset into training and test data subsets. Notably, the Lasso Regression algorithm outperformed the others when evaluating the regression models using performance metrics.