Predicting window opening behavior of individual occupant using machine learning models

Abstract

The purpose of this study was to predict individual occupant window control behavior in a residential building using a machine learning model. Outdoor and indoor environmental conditions and window states of 23 sample housing units were measured every 10 minutes for 10 months. The occupants showed different window opening behavior even under identical environmental conditions. Three machine learning models, k-Nearest Neighbors (KNN), Random Forest (RF) and Artificial Neural Networks (ANN) were used to predict window states of 23 individual occupants. The results shows that machine learning methods are appropriate to predict occupants' individual window opening behavior. © 2018 15th Conference of the International Society of Indoor Air Quality and Climate, INDOOR AIR 2018. All rights reserved.