Utility of Geographic Adjacency
Background: Spatiotemporal epidemic forecasting models often rely on geographic adjacency to capture regional transmission dynamics.
Question / Future Work: It is unclear why commonly used geographic adjacency structures frequently underperform relative to univariate baselines in epidemic forecasting, suggesting that simple geographic proximity may be an ineffective or misleading proxy for underlying epidemiological transmission networks.
Why It Matters: This is a fundamental failure mode in the field that limits the effectiveness of current spatiotemporal modeling approaches in public health.
Evidence: Consistent with statistical epidemiology findings, we conclude that geographic adjacency may be a poor proxy for cross-region dependencies and can hinder performance.
Metadata & Links
- created_at
- 2026-05-10T06:04:03Z