Vehicle CO2 Emission Predictive Analytics Using HistGradientBoosting Regression Algorithms

Abstract

Vehicle CO2 emissions are a significant contributor to climate change, so research on this subject is needed. Strong prediction models and data analysis techniques are required to obtain accurate results. This research aims to analyze and predict vehicle CO2 emissions using machine learning algorithms. Given its efficiency in handling large datasets, the HistGradientBoosting Regression algorithm was selected for predicting vehicle CO2 emissions. The process commenced with meticulous data preparation, which involved cleaning and feature engineering. Key factors such as engine size, fuel economy, and vehicle weight were analyzed to gain insights into their impact on emissions. The study utilized a dataset comprising vehicle specifications and emissions, training and testing the HistGradientBoosting. The model's performance was evaluated using metrics like Mean Absolute Error (MAE),  Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and the Coefficient of Determination (R²). The findings indicate that this approach effectively identifies significant factors influencing emissions while achieving impressive prediction accuracy. This research offers valuable insights for policymakers and manufacturers aiming to develop low-emission vehicles and promote sustainable transportation initiatives. The paper highlights the capability of machine learning to address environmental challenges.