Energy Performance Forecasting in Buildings Using Machine Learning

Energy Performance Forecasting in Buildings Using Machine Learning Algorithms, addresses the critical issue of energy consumption and efficiency in buildings. The primary problem it seeks to overcome is the limitation of existing approaches that provide predictions for the entire building's energy consumption without offering insights into room-wise energy patterns. So, it seeks to overcome this drawback by employing various regression techniques to predict energy levels in each room, enabling a more detailed and understanding of energy usage within the building. This study proposes a comprehensive system for predicting energy performance in each room of a building using various machine learning algorithms, including Linear Regression, Ridge Regression, K-Nearest Neighbors (KNN) Regression, Decision Tree Regression, Random Forest Regression, and Gradient Boosting Regression. The system is designed to provide accurate energy consumption forecasts for each room, ultimately contributing to more efficient energy management within the building. Here using of regression is for energy performance forecasting, regression methods are chosen over classification due to the continuous and quantitative nature of energy consumption variables. Regression models, such as Linear Regression and Decision Tree Regression, provide precise predictions of energy usage levels in buildings, allowing for a understanding of consumption patterns. Additionally, this system goes beyond total energy consumption and further refines the predictions by separately forecasting energy usage for appliances and lighting in each room. By breaking down energy consumption into these components, it offers a more detailed perspective on energy efficiency, allowing for precise optimization of individual room operations. The results show that the successful forecasting R-square of the proposed system is around RFR 0.95%.