Abstract
The indoor environment greatly affects health and well-being; enhancinghealth and reducing energy use in these settings is a key research focus. Withadvancing Information and Communication Technology (ICT), recommendationsystems and reinforcement learning have emerged as promising methods to inducebehavioral changes that improve indoor environments and building energyefficiency. This study employs text-mining and Natural Language Processing(NLP) to examine these approaches in building control and occupant interaction.Analyzing approximately 27,000 articles from the ScienceDirect database, wefound extensive use of recommendation systems and reinforcement learning forspace optimization, location recommendations, and personalized controlsuggestions. Despite broad applications, their use in optimizing indoorenvironments and energy efficiency is limited. Traditional recommendationalgorithms are commonly used, but optimizing indoor conditions and energyefficiency often requires advanced machine learning techniques likereinforcement and deep learning. This review highlights the potential forexpanding recommender systems and reinforcement learning applications inbuildings and indoor environments. Areas for innovation include predictivemaintenance, building-related product recommendations, and optimizingenvironments for specific needs like sleep and productivity enhancements basedon user feedback.