Evaluating urban air quality determinants using Fermatean Picture Fuzzy (FPF)-based Multi-Criteria Decision Making (MCDM) and machine learning techniques

Air pollution has emerged as a critical environmental and public health issue in rapidly urbanizing Indian cities. Identifying and prioritizing the determinants influencing urban air quality remains challenging due to spatial variability, uncertainty, and the nonlinear behavior of atmospheric processes. This paper aims to develop an integrated framework combining Fermatean Picture Fuzzy (FPF)-based Multi-Criteria Decision Making (MCDM) and Machine Learning (ML) techniques to evaluate key urban air quality determinants. The proposed framework systematically ranks meteorological, spatial, and anthropogenic factors through uncertainty-aware fuzzy modeling. Subsequently, the prioritized determinants are validated using the Extreme Gradient Boosting (XGBoost) algorithm to examine their contribution to the Air Quality Index (AQI) in Delhi, India. The findings indicate that traffic density, vegetation index, and spatial coordinates are the most influential determinants, significantly improving model generalization capability. However, optimal prediction performance is achieved using the top ten to eleven determinants, incorporating meteorological, seasonal, diurnal, and population density-related variables. The novelty of this study lies in developing a hybrid framework which provides a transparent, data-driven, and robust approach for prioritizing urban air quality determinants, supporting sustainable urban planning and environmental management decision-making.