init_mcmc_params: init_mcmc_params Finds clever starting points for the MCMC to...

Description Usage Arguments Value Author(s) See Also Examples

View source: R/init_mcmc_params.R

Description

init_mcmc_params Finds values of the serial interval distribution parameters, used to initialise the MCMC estimation of the serial interval distribution. Initial values are computed based on the observed mean and standard deviation of the sample from which the parameters are to be estimated.

Usage

1
init_mcmc_params(si_data, dist = c("G", "W", "L", "off1G", "off1W", "off1L"))

Arguments

si_data

the data on dates of symptoms of pairs of infector/infected individuals to be used to estimate the serial interval distribution. This 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

dist

the parametric distribution to use for the serial interval. Should be one of "G" (Gamma), "W" (Weibull), "L" (Lognormal), "off1G" (Gamma shifted by 1), "off1W" (Weibull shifted by 1), or "off1L" (Lognormal shifted by 1).

Value

A vector containing the initial values for the two parameters of the distribution of the serial interval. These are the shape and scale for all but the lognormal distribution, for which it is the meanlog and sdlog.

Author(s)

Anne Cori

See Also

estimate_R

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
## 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")

## get clever initial values for shape and scale of a Gamma distribution
## fitted to the the data MockRotavirus$si_data
clever_init_param <- init_mcmc_params(MockRotavirus$si_data, "G")

## estimate the serial interval from data using a clever starting point for 
## the MCMC chain
SI_fit_clever <- coarseDataTools::dic.fit.mcmc(dat = MockRotavirus$si_data,
                             dist = "G",
                             init.pars = clever_init_param,
                             burnin = 1000,
                             n.samples = 5000)

## estimate the serial interval from data using a random starting point for 
## the MCMC chain
SI_fit_naive <- coarseDataTools::dic.fit.mcmc(dat = MockRotavirus$si_data,
                             dist = "G",
                             burnin = 1000,
                             n.samples = 5000)


## use check_cdt_samples_convergence to check convergence in both situations
converg_diag_clever <- check_cdt_samples_convergence(SI_fit_clever@samples)
converg_diag_naive <- check_cdt_samples_convergence(SI_fit_naive@samples)
converg_diag_clever
converg_diag_naive


## End(Not run)

EpiEstim documentation built on Jan. 7, 2021, 5:10 p.m.