Estimation of Epidemic Contact Network

The entire globe is now under the shade of the COVID-19 pandemic, which adversely affects normal human life and their socializing behavior. To break the spread of disease across society, a scientific approach must be evolved. Major strategies to break this spreading chain are contact tracing and modeling the epidemic diffusion. The pandemic spread can be perfectly modeled in the vertex domain using a contact network graph. Contact network enables to identify those contacts who are more susceptible to disease. This research work aims to learn contact network using the Gaussian Markov Random field (GMRF) model and Maximum Likeihood (ML) estimation. Synthetic data is generated using Susceptible-Infected-Recovered (SIR) and Susceptible-Exposed-Infected-Recovered (SEIR) mathematical models. Optimized the results using Alternating Direction Method of Multipliers (ADMM), Combinatorial Graph Laplacian (CGL), Majorization-Minimization (MM) and Convex Optimization(CVX). Results show that, among the optimization algorithms adopted, CVX outperforms the others.