Using Machine Learning Techniques to Moderate CO<sub>2</sub> Emission in KSA Future Smart Cities
Arwa H. Alshanbari, Shahd H. Altalhi, Samaher S. Alsharif, Ghalia M. Alharthi, Razan W. Althubiti, Hossam Meshref
Abstract
As the world grapples with the escalating challenge of climate change, the role of innovative technologies in mitigating environmental impacts has never been more important. This study is inspired by the pioneering smart city of NEOM, Saudi Arabia; one of the forefronts of sustainable urban development, leveraging machine learning (ML) techniques to predict and mitigate carbon dioxide (CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>) emissions in the cities of Saudi Arabia. These cities are striving towards NEOM's approach to achieving zero emissions to the greatest extent possible, because NEOM needs the surrounding cities to have low CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emission as well to achieve its goal. This research aims to create models based on machine learning that can precisely predict CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions, utilizing a detailed dataset of vehicle emissions from Kaggle. We apply the following ML models: Logistic Regression, Gradient Boosting, AdaBoost, Bagging, Backpropagation Neural Networks, Naive Bayes, and Support Vector Machines (SVM). The models are evaluated based on their Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>) as well as Confusion Matrix in predicting CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions. Our analysis achieved the highest performance with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> value of 0.99. Gradient Boosting and Bagging achieved the highest accuracies of 99%. Through a comprehensive review of recent studies and the application of diverse machine learning approaches, this proposed research identifies key challenges in reducing CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions in smart cities, emphasizing the critical role of ML in enhancing the accuracy of forecasts across various sectors. Ultimately, that will support policy makers to make informed decisions to sustain a better urban development.