Machine Learning-Based System for Managing Energy Efficiency of Public Buildings: An Approach towards Smart Cities

Abstract

The issue of energy efficiency cannot be overemphasized in the context of smart cities because buildings are the largest consumers of electrical energy, especially public buildings. Although there have been recent developments with the application of machine learning in this domain, there is still a need for improvement in the performance of already deployed applications. This paper aims to proffer a solution to the question of how to improve the state-of-the-art solutions of machine learning-based intelligent systems for managing the energy efficiency of public buildings as a substantial part of the smart city concept. Gradient Boosting Machine algorithms which include Light Gradient Boosting Machine (LightGBM), Category Boosting (CatBoost), and Adaptive Boosting (AdaBoost) were used to create predictive models of energy consumption of specific public buildings. The best performing model with the minimum root mean squared error of 1.119 was produced by the Light Gradient Boosting Machine algorithm (LightGBM), and a comparison of important predictors extracted by all the methods has been conducted. The evaluation metrics used in this study are Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Coefficient of Determination (R squared). The model could be deployed as an intelligent machine learning-based energy management system for public buildings which could be used as a part of the smart city concept such that public administration could use it to plan reconstruction measures of public buildings to reduce energy consumption and cost. This kind of digital transformation in energy management can further increase the energy efficiency of public buildings and also promote a healthier environment.