A Knowledge Graph and Graph Neural Network Framework for Air Quality-Health Relationships

Elisa Drouot, Thierno Diallo, Gayo Diallo

Keywords: Indoor Air quality, Health Effects, Link Prediction, Knowledge Graph, GNN

Abstract: The air we breathe comes predominantly from indoor environments, where we spend an estimated 90% of our time. As indoor air quality represents a major public health concern, numerous studies have established significant links between exposure to various pollutants and human health. In this context, our research aims to integrate heterogeneous temporal environmental, from buildings, toxicogenomics, and aggregated health records (such as disease prevalence and causes of death). By employing an approach based on Knowledge Graphs and Graph Neural Networks (GNN), we seek to characterize and predict potential associations between pollutants and pathologies. Preliminary results demonstrate the ability to achieve a training loss of 0.0942 and a validation AUC (Area Under the Curve) of 0.8473.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 11