Machine Learning & Uncertainty Quantification: Application in Building Energy Consumption

One of the main tasks in reliability and risk assessment of complex systems is to predict system behavior under uncertainty. Different machine learning (ML) methods can be used to predict system behavior; these methods are able to predict system performance/status in different operating and environmental conditions. In addition to predicting the system's performance, the uncertainty level of the output is important for reliability and risk assessment methods. In this paper, a method for uncertainty quantification of machine learning-based prediction methods is applied on real world data.