ABSTRACT. Residential buildings contribute to the global greenhouse gas (GHG’s) emissions. Besides in current energy efficient buildings, the energy consumption becomes very sensitive to the occupants behaviour. Multiple systems of human behaviour modelling have been designed, to help occupants to reduce their energy consumption. In this paper it is used a system which is based on interactive and cooperative learning for activity and occupancy recognition; adding a new characteristic. The added property allows the system to use detailed and accurate information to model human behavior in the different, complex and multiple scenarios of the residential buildings. Besides, the occupants are able to answer questions of their interest that are related to their energy consumption and comfort impact by creating what in this project are called “experiments”. The latter consist on the freedom of sensors selection and activity impact to be analyzed by the system.