Tree-Based Machine Learning Models for Respiratory Health Prediction in the UAE: Integration of Air Quality and Meteorological Data

Abstract

Purpose: This study aims to develop a data-driven predictive framework to assess and forecast respiratory morbidity risks associated with air pollution and climate variability across the UAE, addressing the gap in region-specific health-environment models. Design/Methodology/Approach: Using environmental data from the National Centre of Meteorology (NCM) and health data from the Ministry of Health and Prevention (MOHAP), machine learning models, namely Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Long Short-Term Memory (LSTM), will be applied to analyse time-series data integrating air pollutants, namely Particulate Matter ≤ 2.5 μm (PM₂.₅), Particulate Matter ≤ 10 μm (PM₁₀), Nitrogen Dioxide (NO₂), Sulphur Dioxide (SO₂), Ozone (O₃), Carbon Monoxide (CO), together with meteorological factors (temperature, humidity, wind speed). Findings: The expected outcome is an explainable AI system capable of issuing early warnings for high-risk respiratory health days, improving predictive accuracy through multi-factor integration. Original/Value of the Paper: The study contributes an innovative, UAE-specific, explainable AI framework that supports public health resilience, hospital preparedness, and environmental policy under UAE Vision 2031. Keywords: Air Pollution; Machine Learning; Respiratory Health; Predictive Modelling; United Arab Emirates; Explainable Artificial Intelligence; Time-Series Forecasting; Environmental Health Policy Citation: Mohamad, M., Ivanova, G., Kohli, A. and Srivel, R. (2026): Tree-Based Machine Learning Models for Respiratory Health Prediction in the UAE: Integration of Air Quality and Meteorological Data. World Journal of Science, Technology and Sustainable Development, Vol. 21, No. 6, pp. 545-xx