Consistent Approximation of Epidemic Dynamics on Degree-heterogeneous Clustered Networks
arXiv:1810.06066
Abstract
Realistic human contact networks capable of spreading infectious disease, for example studied in social contact surveys, exhibit both significant degree heterogeneity and clustering, both of which greatly affect epidemic dynamics. To understand the joint effects of these two network properties on epidemic dynamics, the effective degree model of Lindquist et al. is reformulated with a new moment closure to apply to highly clustered networks. A simulation study comparing alternative ODE models and stochastic simulations is performed for SIR (Susceptible-Infected-Removed) epidemic dynamics, including a test for the conjectured error behaviour in Pellis et al., providing evidence that this novel model can be a more accurate approximation to epidemic dynamics on complex networks than existing approaches.
arXiv comment14 pages, 7 figures
References in corpus (10)
- Epidemic processes in complex networks
- Random graphs with clustering
- Properties of highly clustered networks
- Edge-Based Compartmental Modeling for Infectious Disease Spread Part I: An Overview
- Random graphs containing arbitrary distributions of subgraphs
- Percolation and Epidemic Thresholds in Clustered Networks
- Efficient and exact sampling of simple graphs with given arbitrary degree sequence
- Bond percolation on a class of clustered random networks
- Maximum-entropy moment-closure for stochastic systems on networks
- Endemic infections are always possible on regular networks