This study presents an advanced Maintenance 4.0 framework designed for the systematic assessment and mitigation of indoor air quality (IAQ) risks within university laboratory settings. The framework leverages real‐time monitoring via IQAir AirVisual Pro sensors to continuously track critical IAQ parameters, specifically particulate matter (PM 1 , PM 2.5 , and PM 10 ), carbon dioxide (CO 2 ), temperature, and relative humidity. To facilitate a shift from reactive to predictive maintenance strategies, the research integrates machine learning–based predictive models with statistical time‐series forecasting. Among the evaluated architectures, the stacking ensemble model achieved superior performance in predicting the Air Quality Index (AQI), yielding a root mean square error (RMSE) of 3.06, a mean absolute error (MAE) of 1.61, and an R 2 of 0.97, thereby outperforming random forest, k ‐nearest neighbor, and gradient boosting algorithms. Concurrently, the vector autoregression (VAR) model demonstrated robust short‐term predictive capability (RMSE 2.93, MAE 0.98, and R 2 0.97), confirming its effectiveness in capturing temporal AQI dynamics and providing an interpretable baseline. The synergistic application of VAR and machine learning enhances prediction robustness, anomaly detection, and maintenance decision support. By enabling timely, data‐driven interventions, the framework aligns with international standards, including ASHRAE 62.1, the World Health Organization (WHO), and the U.S. Environmental Protection Agency (EPA), offering a scalable methodology for ensuring safer and more sustainable laboratory environments.
Abdullah et al. (Thu,) studied this question.