NRG4-PRO: A NOVEL FULL INFERENCE ENGINE

The challenges in energy industry are rising both for the energy providers and the consumers. This paper proposes a novel tool (NRG4-Pro) for energy analytics that creates a personal energy profile for each user without requiring ground information. Indoor comfort levels, user occupancy and activity inference are estimated and analyzed in real-time. A novel fusion algorithm is proposed for thermal and visual comfort. Users and device energy baseline are inferred using Hidden Markov Models in an unsupervised manner, while occupancy and activity are inferred utilizing Decision Trees and Random Forest Classifiers. The proposed energy analytics tool (NRG4-Pro) has been thoroughly tested in residential environments, while the experimental results shows that it is a powerful tool for energy profiling and indoor environment awareness.