Boundary Degree as a Node-level Feature for Epidemic Scenario Identification in Agent-based Cascade Simulations
For researchers using agent-based simulations to analyze epidemics, this work provides a simple yet effective feature that improves scenario identification, with practical implications for contact tracing applications.
The paper proposes boundary degree as a per-node cascade feature for identifying epidemic scenarios from disease cascades, showing a 19% improvement in accuracy on realistic social contact networks. It also provides theoretical grounding for the importance of edge features and proves that certain scenarios are indistinguishable without boundary or edge information.
Characterizing the scenario underlying an epidemic from its disease cascade is an important task in simulation analytics. We propose boundary degree, the count of an infected node's contacts in the underlying contact network that were not infected, as a per-node cascade feature for this task. Through systematic ablation on realistic social contact networks of Tennessee and Virginia, we show that boundary degree alone improves scenario identification accuracy by 19%. Edge features, whose importance was observed empirically by prior work, consistently improve accuracy across all settings; we provide theoretical grounding for this observation. These effects are complementary. We prove that certain epidemic scenarios are indistinguishable without boundary or edge information. Prior feature engineering approaches included aggregate boundary statistics, but these were not among the top-ranked feature groups; the per-node representation we propose reveals their importance clearly. Our results suggest that contact tracing applications should track contacts with non-infected individuals, not only transmissions.