Machine Learning for Indoor Air Quality Assessment: A Systematic Review and Analysis

The health and well-being of people are highly influenced by the quality of air they breathe in. Since human beings spend the majority of their time indoors, the assessment of indoor air quality (IAQ) is a critical concern. As per the reports from the World Health Organization, degraded air quality leads to a massive burden of mortality and morbidity in middle- and low-income countries. Therefore, timely assessment and forecasting of critical scenarios are essential to avoid the onsets of chronic problems like respiratory illness, cardiovascular disease, cancer, and degenerative disorders. This systematic literature review focuses on the potential of machine learning methods for IAQ assessment. This work includes 18 studies extracted from four different databases (IEEE Explore, Web of Science, Scopus, and PubMed) from year 2020 to 2022. The results show that majority of studies preferred using Random Forest for IAQ parameter forecasting and Root Mean Squared Error is preferred as primary evaluation metrics for model performance assessment. Furthermore, PM2.5 has been considered a potential cause of critical IAQ conditions among several other pollutants such as CO2, PM10, VOC, CO, and NO2. This work provides answers to seven research questions in the IAQ assessment domain and exposes potential opportunities, challenges, and recommendations for future researchers.