FORECASTING ELECTRICITY CONSUMPTION OF RESIDENTIAL USERS BASED ON LIFESTYLE DATA USING ARTIFICIAL NEURAL NETWORKS

Electricity is the lifeline of almost everything in this 21st century. Residential electricity consumption has seen an increase both locally and globally. Therefore, it has become a global concern of significant importance to promote electrical energy consumption reduction (energy conservation) within the household for a viable development of a nation in the case of resource limitations. The current study seeks to identify the social psychology (lifestyle) factors that significantly influence the residential electricity consumption, and predict future electricity consumption using an artificial neural network (ANN) based on lifestyle data collected from three hundred and fifty (350) households in the Sunyani Municipality. The performance metrics RMSE, MSE, MAPE, and MAE, were used to estimate the performance of the proposed model. The RMSE (0.000726) and MAE (0.000976) of the proposed model compared to (RMSE = 0.0657 and MAE = 0.05714) for Decision Trees (DT) and (RMSE = 0.08816 and MAE = 0.06911) for Support Vector Regression (SVR) shows a better fit of the proposed model. Furthermore, it was observed that the type of vehicle (saloon or sport utility vehicle) used by the head of a household was the most significant lifestyle feature in forecasting residential electricity consumption. Future studies would focus on developing a vigorous model using a combination of weather parameters and several socioeconomic factors based on hybrid machine-learning algorithms to increase forecasting accuracy.