Authors
Angelo Casolaro, Vincenzo Capone, Massimiliano Giordano Orsini, Francesco Camastra
Conference
Abstract
Environmental spatiotemporal prediction is crucial for air quality management, where accurate prediction of primary air pollutants is essential for public health and policymaking. This paper introduces the Conditioned Diffusion-based Graph Attention (CDGA) model, a novel Bayesian deep learning framework for spatiotemporal prediction of primary air pollutant ground-level concentrations. CDGA integrates Graph Attention Networks (GAT) and time series model order as conditioning inputs, enabling the model to capture both spatial and temporal dependencies while it provides uncertainty measures for the predictions. The proposed model is validated on two spatiotemporal benchmarks of NO2 and O3 ground-level concentrations, measured by EEA stations in Italy from 2014 to 2022. Experimental results demonstrate that CDGA outperforms state-of-the-art spatiotemporal models in terms of popular error metrics.
Description
The paper “Environmental Spatiotemporal Prediction with a Conditioned Diffusion-based Graph Attention Model” introduces an innovative Bayesian deep learning framework for forecasting air pollutant concentrations.
The proposed Conditioned Diffusion-based Graph Attention (CDGA) model combines Graph Attention Networks with time series conditioning to effectively capture spatial and temporal dependencies in environmental data.
Applied to NO₂ and O₃ measurements collected across Italy between 2014 and 2022, CDGA demonstrated superior predictive accuracy and robust uncertainty estimation compared to existing spatiotemporal models. This approach represents a significant step toward more reliable air quality forecasting and data-driven environmental decision-making.