Conference Paper

Proceedings of Building Simulation 2019: 16th Conference of IBPSA

Study on Occupancy Prediction for Building Operation using Machine Learning Method

Yuan Jin, Da Yan, Hongsan Sun
Tsinghua University, China, People's Republic of

DOI: https://doi.org/10.26868/25222708.2019.210370
Abstract: Occupancy in buildings is an important input no matter to building design or building on-site operation. As for building design, the occupancy state is related to the heat gain in buildings and further equipment operation. With room occupied, the comfortable ambient parameter is required and the equipment in buildings are potential to be used. As for the building operation, with the detection of occupancy, the auto control is likely to perform. However, the accuracy needs improved. In this study, occupancy state is predicted by proposed machine learning method. During the occupancy prediction, the temporal sequential characteristic is considered to obtain a better prediction performance. The evaluation indicator is defined and the evaluation is conducted through ten fold cross validation. This study can be further promoted to building operation and performance simulation.
Keywords: occupancy, cooling and heating load, human behavior, simulation
Pages: 2157 - 2164
Paper: