Automated energy consumption forecasting with EnForce

Download PDF

Abstract

The need to reduce energy consumption on a global scale has beenof high importance during the last years. Research has createdmethods to make highly accurate forecasts on the energy consump-tion of buildings and there have been efforts towards the provisionof automated forecasting for time series prediction problems. En-Force is a novel system that provides fully automatic forecastingon time series data, referring to the energy consumption of build-ings. It uses statistical techniques and deep learning methods tomake predictions on univariate or multivariate time series data,so that exogenous factors, such as outside temperature, are takeninto account. Moreover, the proposed system provides automaticdata preprocessing and, therefore, handles noisy data, with missingvalues and outliers. EnForce includes full API support and can beused both by experts and non-experts. The proposed demonstrationshowcases the advantages and technical features of EnForce.