A machine-learning based approach to model user occupancy and activity patterns for energy saving in buildings

摘要: Recently it has been noted that user behaviour can have a large impact on the final energy consumption in buildings. Through combination of mathematical modelling and data from wireless ambient sensors, we model human patterns use information to regulate building management systems (BMS) order achieve best trade-off between comfort efficiency. In this work, modelled occupancy activity using Machine Learning approaches. We applied non-linear multiclass Support Vector Machines (SVMs) deal with complex nature collected various sensors accurately identify activities daily living (ADL) patterns. To validate our results, also used other methodologies (i.e. Hidden-Markov Model k-Nearest Neighbours). The experimental results show proposed approach outperforms methods for scenarios evaluated.