Learn R Programming

EpiEstim (version 2.2-4)

make_mcmc_control: make_mcmc_control Creates a list of mcmc control parameters to be used in config$mcmc_control, where config is an argument of the estimate_R function. This is used to configure the MCMC chain used to estimate the serial interval within estimate_R (with method "si_from_data").

Description

make_mcmc_control Creates a list of mcmc control parameters to be used in config$mcmc_control, where config is an argument of the estimate_R function. This is used to configure the MCMC chain used to estimate the serial interval within estimate_R (with method "si_from_data").

Usage

make_mcmc_control(
  burnin = 3000,
  thin = 10,
  seed = as.integer(Sys.time()),
  init_pars = NULL
)

Arguments

burnin

A positive integer giving the burnin used in the MCMC when estimating the serial interval distribution.

thin

A positive integer corresponding to thinning parameter; the MCMC will be run for burnin+n1*thin iterations; 1 in thin iterations will be recorded, after the burnin phase, so the posterior sample size is n1.

seed

An integer used as the seed for the random number generator at the start of the MCMC estimation; useful to get reproducible results.

init_pars

vector of size 2 corresponding to the initial values of parameters to use for the SI distribution. This is the shape and scale for all but the lognormal distribution, for which it is the meanlog and sdlog.

Value

An object of class estimate_R_mcmc_control with components burnin, thin, seed, init_pars. This can be used as an argument of function make_config.

Details

The argument si_data, should be a dataframe with 5 columns:

  • EL: the lower bound of the symptom onset date of the infector (given as an integer)

  • ER: the upper bound of the symptom onset date of the infector (given as an integer). Should be such that ER>=EL

  • SL: the lower bound of the symptom onset date of the infected individual (given as an integer)

  • SR: the upper bound of the symptom onset date of the infected individual (given as an integer). Should be such that SR>=SL

  • type (optional): can have entries 0, 1, or 2, corresponding to doubly interval-censored, single interval-censored or exact observations, respectively, see Reich et al. Statist. Med. 2009. If not specified, this will be automatically computed from the dates

Assuming a given parametric distribution for the serial interval distribution (specified in si_parametric_distr), the posterior distribution of the serial interval is estimated directly fom these data using MCMC methods implemented in the package

Examples

Run this code
# NOT RUN {
## Note the following examples use an MCMC routine
## to estimate the serial interval distribution from data,
## so they may take a few minutes to run

## load data on rotavirus
data("MockRotavirus")

## estimate the reproduction number (method "si_from_data")
mcmc_seed <- 1
burnin <- 1000
thin <- 10
mcmc_control <- make_mcmc_control(burnin = burnin, thin = thin, 
                     seed = mcmc_seed)

incid <- MockRotavirus$incidence
method <- "si_from_data"
overall_seed <- 2
config <- make_config(incid = incid, 
                     method = method, 
                     si_parametric_distr = "G",
                     mcmc_control = mcmc_control,
                     n1 = 500
                     n2 = 50,
                     seed = overall_seed)

R_si_from_data <- estimate_R(incid,
                            method = method,
                            si_data = MockRotavirus$si_data,
                            config = config)
# }

Run the code above in your browser using DataLab