Epidemic modelling at a community level using contact networks

Abstract

In this thesis, an epidemic modelling problem at an individual level is studied. The Individual-Based model, which is the discretised version of the SIR compartmental model on a network, is used to study the problem. The objectives are to recover the infected rate assumed in the ground truth and explore testing strategies to mitigate the spread quickly. A maximum likelihood estimator is used for inference of the parameter. The probabilities for the MLE are derived using the stochastic version of the IB model called the Individual-Based Monte Carlo (IB-MC) model. The synthetic data and two real-world data sets used for modelling are described in detail. An edge-centric Contact-Based model is also discussed for temporal networks. The differences between the two models and the difference in the simulation models and the mean-field models are explored. Testing strategies that vary in both time and selection of individuals are presented. Numerical results from the two real-world data sets are presented to investigate the accuracy of the estimated parameter from the testing strategies proposed.