Towards a Better Estimation of Indoor Air Quality (IAQ): An Approach based on Quality of Experience (QoE)

Indoor air quality is a major health and well-being concern, particularly in educational settings where children spend a large part of their day. Monitoring parameters such as carbon dioxide (CO₂) and fine particulate matter (PM₂.₅) helps identify instances of poor air quality; however, interpreting these measurements often remains complex for users and facility managers. Furthermore, existing recommendations are based on a variety of different standards and thresholds. In this work, we propose a new approach inspired by the concept of Quality of Experience (QoE)—commonly used in digital and multimedia services to measure quality as perceived by users. We adapt the Mean Opinion Score (MOS) principle, typically defined on a scale from 1 to 5, to the assessment of indoor air quality. The proposed indicator, MOS-IAQ (Mean Opinion Score – Indoor Air Quality), integrates recommendations from the World Health Organization (WHO), the U.S. Environmental Protection Agency (EPA), and ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) to create a unified, five-level scale, ranging from "Excellent" quality to a "Dangerous" situation. This approach is tested using four public datasets from primary school classrooms located in various countries around the world (including European nations, Madagascar, South Africa, and Mexico). Several machine learning techniques were compared for their ability to predict the MOS-IAQ level based on environmental measurements. The results show that the Random Forest algorithm, in particular, achieves an accuracy of approximately 99.99%. This approach paves the way for simple, real-time monitoring and prediction systems, enabling complex environmental data to be translated into easily understandable information regarding the level of risk to occupants. Finally, we believe that this simple and interpretable approach will play an important role in improving indoor air quality monitoring in schools and could be adapted for other environments, such as offices, hospitals, or residential settings.