Direct and Surrogate Likelihood-Free Statistical Inference for Epidemiological Models in a Network of Contacts
Modeling the dynamic of epidemic outbreaks in a network can help to describe not only infectious events in a population but also other types of infectious agents, such as information and ideas. The pairwise contact network structure in the population is introduced to relax the mass action assumption, but the model increases its complexity and in most cases it is extremely expensive, or impossible, obtaining its likelihood function. In this work we propose the statistical inference of epidemic compartmental models based on the Approximate Bayesian Computation (ABC) and propose a surrogate version to reduce the required computer time. We illustrate this proposal with computational experiments of epidemic outbreaks in simulated networks.
KeywordsModelling approach Evaluation Index System AHP
- 1.Allen, L.J.: An introduction to stochastic epidemic models. In: Mathematical Epidemiology, pp. 81–130. Springer, Berlin (2008)Google Scholar
- 5.Dutta, R., Mira, A., Onnela, J.P.: Bayesian inference of spreading processes on networks. In: Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 474(2215), p. 20180,129 (2018)Google Scholar
- 7.Gillespie JH (1976) A general model to account for enzyme variation in natural populations. ii. characterization of the fitness functions. Am. Nat. 110(975), 809–821Google Scholar
- 8.Kiss, I.Z., Miller, J.C., Simon, P.L.: Mathematics of Epidemics on Networks. Springer, Berlin (2017)Google Scholar
- 10.Marjoram P, Molitor J, Plagnol V, Tavaré S (2003) Markov chain monte carlo without likelihoods. In: Proceedings of the National Academy of Sciences 100(26), pp. 15,324–15,328Google Scholar