Auto-Configured Explainable Graph Neural Networks for Multi-Site Pollution Prediction

2026-06-23

AI Summary

This study addresses the limitations of conventional graph neural networks (GNNs), which rely on fixed graph structures and thus struggle to capture the dynamic spatiotemporal dependencies inherent in multi-site air pollution data while offering limited interpretability. To overcome these challenges, the authors propose a novel approach that dynamically constructs graph structures using supervised learning based on confusion matrices. The method employs a hybrid loss function combining energy distance and Huber loss to mitigate gradient vanishing and enhance prediction stability. Furthermore, model interpretability is improved through integration with GNNExplainer and PGExplainer. Evaluated on PM1, PM2.5, and PM10 concentration forecasting tasks, the proposed framework—built upon GraphSAGE as the backbone—consistently outperforms baseline models such as Prophet, LSTM, and GRU in both single-step and multi-step predictions, achieving a unified balance of high accuracy, robustness, and interpretability.

Abstract

Accurate particulate matter (PM) prediction is crucial for mitigating air pollution. Graph Neural Networks (GNNs) effectively model spatiotemporal dependencies, but predefined graphs limit adaptability, and some datasets complicate learning. This study introduces a graph construction method based on a confusion matrix from a supervised learning process to dynamically capture inter-class relationships. Additionally, a hybrid loss function that combines energy distance and Huber loss is applied to address the vanishing gradient problem and improve learning stability. The approach is evaluated using air pollution data from the University of Utah AirU Pollution Monitoring Network in Salt Lake City, UT, with five GNN models: Graph Convolutional Networks (GCNs), Simple Graph Convolutional Networks (SGConv), Graph Isomorphism Networks (GINs), Graph Attention Networks (GATs), and GraphSage. The experimental results of single- and multistep predictions confirm that GraphSage achieves the highest accuracy in predicting the concentrations of PM${1}$, PM${10}$, and PM$_{2.5}$ over different time horizons. Furthermore, {\color{black} GNNExplainer (Graph Neural Network Explainer) and PGExplainer (Probabilistic Graph Explainer)} are applied to interpret feature importance and graph structure, ensuring model transparency. Results show improved prediction accuracy, with GNN models outperforming traditional machine learning \textcolor{black}{and deep learning models (i.e., Prophet, Long short-term memory, Gated recurrent units} in air pollution forecasting.

Problem

• Graph Neural Networks
• Air Pollution Prediction
• Spatiotemporal Dependencies
• Multi-Site Prediction
• Particulate Matter

Innovation

• dynamic graph construction
• hybrid loss function
• explainable GNN
• spatiotemporal pollution prediction
• confusion matrix-based graph