Ensemble machine learning framework for daylight modelling of various building layouts

Abstract The application of machine learning (ML) modelling in daylight prediction has been a promising approach for reliable and effective visual comfort assessment. Although many advancements have made, no standardized ML framework exists In this study, 625 different building layouts were generated to model useful illuminance (UDI). Two state-of-the-art algorithms, eXtreme Gradient Boosting (XGBoost) random forest (RF), employed analyze UDI four categories: UDI- f (fell short), s (supplementary), (autonomous), e (exceeded). A feature (internal finish) was introduced the better reflect real-world representation. results show that XGBoost models predict with maximum accuracy R 2 = 0.992. Compared RF, can significantly reduce errors. Future research directions specified advance proposed by introducing new features exploring architectures standardize applications prediction.