Exploring the Predictive Power of Machine Learning for Energy Consumption in Buildings

Abstract

This study investigates the application of machine learning techniques to predict energy consumption in buildings. Five machine learning models (linear regression, decision trees, random forest, support vector machines, and neural networks) are evaluated, and a permutation feature importance analysis is conducted to identify the most influential features. The random forest model emerges as the top performer, achieving a mean absolute error of 15.2%. Temperature, solar radiation, and relative humidity are found to be the important features in energy prediction. The study demonstrates the potential of machine learning for energy prediction in buildings, contributing to more efficient energy management and sustainability.