autoInitConv: Run MCMC iterations until initial convergence
In thlytras/FluHMM: Hidden Markov Model for influenza sentinel surveillance

Description Usage Arguments Details Value

This function is not normally called by the user; it is called from FluHMM provided initConv=TRUE (the default). It generates posterior samples from the model repeatedly until convergence is reached for the sigma[1] parameter (this is called "initial convergence").

1	autoInitConv(x, iter = 5000, maxit = 95000, Rhat = 1.1)

`x`	An object of class ‘FluHMM’, from which to generate posterior samples.
`iter`	Number of iterations to run.
`maxit`	Maximum number of iterations to run before giving up.
`Rhat`	Gelman-Rubin diagnostic cutoff value to determine convergence (only for the sigma[1] parameter).

The sigma[1] parameter (standard deviation of the pre-epidemic phase) is of primary importance in the model, since the pre-epidemic phase comes first and its correct identification is the basis on which to estimate the other phases. If convergence for sigma[1] has been reached, the other parameters in the model are very likely to have converged too, with the exception of beta[2] and beta[3] (slopes of the epidemic growth and epidemic decline phase); the latter mix more slowly and may necessitate a longer, and possibly thinned, MCMC chain.

Therefore "initial convergence" is defined as convergence for the sigma[1] parameter. Unless this is achieved, it is inadvisable to use the posterior samples for any inference at all. Only _after_ this has been achieved can a new posterior sample be generated (using update). Then convergence for all parameters is checked again and, if not achieved, a new sample is generated from scratch or the current one further is further extended.

None. The function mutates its argument ‘x’ directly.

thlytras/FluHMM documentation built on May 31, 2019, 10:44 a.m.