Work

Mobility vs. Contiguity: Spatially Explicit Graph Neural Networks for COVID-19 Forecasting

Year: 2025

Type: conference-paper

Abstract: This study demonstrates that graph construction is a critical modelling decision in spatial GNN forecasting by introducing and systematically evaluating five alternative spatial graph designs. The findings show that selecting or combining graph structures to reflect local mobility and spatial configurations can substantially improve predictive performance, underscoring the need to tailor graph representations to urban context.

Source: ENLIGHTEN (Jurnal Bimbingan dan Konseling Islam)

DOI: 10.5281/zenodo.17660771

Authors: Li, Fengjiao, Wu, Meiliu, Basiri, Anahid

Institution: University of Glasgow

Language: English

Topic: COVID-19 epidemiological studies

Subfield: Modeling and Simulation

Field: Mathematics

Domain: Physical Sciences

SDG: Sustainable cities and communities

Open Access status: green

Funders: University of Glasgow, China Scholarship Council