A prediction system for home appliance usage

► The prediction of appliance usage in smart homes is done, as an important problem in the home energy management. ► A global architecture for prediction with an oracle module and a predictor module was proposed. ► Several classifiers were tested as predictors and for each tested appliance, the best one was detected. ► The prediction with decision tables for lighting usage in dwellings gives the highest accuracy in all the tested cases. ► The oracle module provides additional information which leads to more accurate prediction.