LAUSR

Abstract Occupant behavior is becoming a major factor impacting indoor environments and building operations. Individual differences in indoor environment preferences impact the implementation of occupant-centric building operations. Previous studies regarding occupant behavior have focused on improving the building simulation performance by establishing a representative occupant behavior model to reduce the gap between the real performance and simulation results. In this study, a customized shading behavior identification model (CSBIM) is developed to learn and identify the personalized shading behavior habits and indoor environmental preferences of targeted occupants during the building operation phase. The CSBIM can be used to extract personalized behavioral preferences and embedded into the optimization controls of occupant-centric buildings. The proposed CSBIM is developed based on the cognitive concept of subliminal perception. Monitored data from an office on campus is used to select the optimal classification algorithm and evaluate the effectiveness of the CSBIM. The results indicate that using CSBIM – Random Forest improved capturing the personalized shading behavior by 12.56%. Finally, the advantages and limitations of the CSBIM are also considered.