inf_time_proposal: PhD A Gibbs Sampler for a partially observed SIR epidemic.

Description Usage Details

View source: R/Centered_MCMC.R

Description

A Centered Parameterisation of Bayesian Modelling of the Infection and Removal times of an Epidemic Centered

Usage

1
inf_time_proposal(t_inf, t_rem, no_proposals, par_rem)

Details

==== Model ====

theta(beta, gamma) –> Infection Times (I, Unobserved) –> Removal Times (R, Observed)

Approach is based on that used by Neal & Roberts (2003).

1. Take observed data and simulate a set of valid infection times using inital values for gamma and beta.

Steps for one iteration:

2. Calculate the parameters of the conditional posteriors to be sampled from.

3. Sample from the conditional distributions for the infection and removal rate in turn

4. Sample an Infection Time to change. Propose the change and Accept/Reject using Metropolis-Hastings Acceptance probability


JMacDonaldPhD/Epidemics documentation built on Jan. 10, 2020, 2:48 a.m.