A Human–Building Interaction Framework for Feedback Curation and Predictive Modeling in Occupant-Centric Indoor Environments

Occupant-centric indoor environment management requires consistent occupant feedback to optimize indoor environmental quality (IEQ) and well-being. However, existing systems often fail to collect and integrate multidimensional feedback with environmental data, limiting their ability to provide actionable insights. This paper proposes a novel human–building interaction (HBI) framework that combines occupant feedback collection user interface (UI), environmental monitoring, and an advanced deep learning-based predictive model. The framework collects occupant feedback on comfort, health, and emotion levels; integrates it with IEQ and outdoor environmental data; and utilizes a model fusion-based LSTM with cross-modality attention to predict feedback trends. A pilot study in an office environment validated the framework, showing its ability to capture frequent and reliable feedback under diverse conditions, and outperformed several baseline deep learning models. This paper highlights the potential of occupant feedback-integrated framework for adaptive IEQ management and lays the foundation for future applications in occupant-centric buildings.