Towards Context-aware Power Forecasting in Smart-homes

Forecasting future power consumption in residential buildings is important to optimize the grid, assist inhabitants everyday activities, and save energy. Several machine learning methods have been proposed predict electricity smart homes based on history of past data acquired from meters. However, increasing availability home sensors can provide insights about routines activities inhabitants, that may be exploited more accurate predictions. In this paper, we propose a approach forecast energy considering not only data, but also context such as inhabitants’ actions use household appliances, interaction with furniture doors, environmental data. We performed an experimental evaluation real-world instrumented environment large set users. The results comparison two baseline show our promising.