Estimates of residential building energy consumption using a multi-verse optimizer-based support vector machine with k-fold cross-validation

The ever-increasing human population, building constructions, and technology usages have currently caused electric consumption to grow significantly. Accordingly, some of the efficient tools for more and more energy saving and development are efficient energy management and forecasting energy consumption for buildings. Additionally, efficient energy management and smart restructuring can improve energy performance in different areas. Given that electricity is the main form of energy that is consumed in residential buildings, forecasting the electrical energy consumption in a building will bring significant benefits to the building and business owners. All these means call for precise energy forecast to make the best decisions. In recent years, artificial intelligence, generally, and machine learning methods, in some areas, have been employed to forecast building energy consumption and efficiency. The present study aims to predict energy consumption with higher accuracy and lower run time. We optimize the parameters of a support vector machine (SVM) using a multi-verse optimizer (MVO) without the grid search algorithm, due to the development consequence of residential energy predicting models. This paper presented the MVO-SVM approach for predicting energy consumption in residential buildings. The proposed approach examined a UCI repository dataset. Based on the experimental results MVO can effectively decrease the number of features while preserving a great predicting precision.