Multi-Classification Protocol for Building Electricity Consumption Modelling Using Neural Networks

The growing demand for energy in developed and developing countries has made it necessary to evaluate energy use to conserve energy. Many studies have been conducted to understand the energy use of appliances in buildings. Effective energy management and accurate forecasting of building energy consumption have become essential factors in investment decision-making and increasing the market value of construction projects. This study proposes an ANN multi-classification to optimize energy use in buildings using real data. By using Sobol indices with Ishigami function for assigning weights to input variables, we show their importance relation support to train ANN model. Furthermore, the study results serve as reference data for engineers and managers in the design and management of buildings.