Energy Consumption Forecasting to Small and Medium Enterprises Using a Hybrid Method

In the paper, a hybrid method based on the linear regression model improved with clustering techniques is proposed for the electricity consumption forecasting to Small and Medium Enterprises (SMEs). The method has as its starting point the dependence between electricity consumption and climatic factors (temperature, humidity and wind speed). The testing was made considering the case of a small enterprise and the obtained results led at smaller forecasting errors than in the case of using the simple linear regression method, with the average error in the analysed period (a week) by 2.72 % compared to 4.44 %.