Gaussian process emulators for spatial individual‐level models of infectious disease. Issue 4 (December 2016)

• Record Type: Journal Article
• Authors: Pokharel, Gyanendra; Deardon, Rob
• Is Part Of: Canadian journal of statistics. Volume 44:Issue 4(2016)
• Page Start: 480
• Page End: 501
• Publication Date: 2016-12
• Subjects: Disease models -- emulators -- Gaussian process -- spatial models -- tomato spotted wilt virus -- MSC 2010: Primary 92B15 -- secondary 62F15; Mathematical statistics -- Periodicals; 519.5
• DOI: 10.1002/cjs.11304
• Languages: English
• ISSNs: 0319-5724

Abstract

Abstract: Statistical inference for spatial models of infectious disease spread is often very computationally expensive. These models are generally fitted in a Bayesian Markov chain Monte Carlo (MCMC) framework, which requires multiple iterations of the computationally cumbersome likelihood function. We here propose a method of inference based on so‐called emulation techniques. Once again the method is set in a Bayesian MCMC context, but avoids calculation of the computationally expensive likelihood function by replacing it with a Gaussian process approximation of the likelihood function built from simulated data. We show that such a method can be used to infer the model parameters and underlying characteristics of the spatial disease system, and this can be done in a computationally efficient manner. The Canadian Journal of Statistics 44: 480–501; 2016 © 2016 Statistical Society of Canada