MCMCfit: Markov Chain Monte Carlo Multiple Systems Estimation

In bernardsilverman/modslavmse: Multiple Systems Estimates for estimating the prevalence of Modern Slavery

Description

Use the MCMCpack routine MCMCpoisson to fit a multiple systems estimation model.

Usage

MCMCfit(zdata, priorprec = 1, mainprec = 1e-04, sigthresh = 2,
  totalpopest = F, runzerothresh = NULL, ...)

Arguments

zdata	an incidence matrix of the form supplied to MSEfit
priorprec	If positive, the reciprocal of the prior variance of the two factor interaction parameters. If zero, an improper prior over (-∞, ∞). If negative, then only main effects are fitted.
mainprec	The reciprocal of the prior variance of the main effect parameters. Only used if priorprec > 0.
sigthresh	If sigthresh > 0 the analysis is repeated dropping out any effects whose ratio of posterior mean to standard deviation does not reach sigthresh
totalpopest	If totalpopest=T then routine outputs the MCMC of the dark figure plus the original data, in other words an estimate of the total population
runzerothresh	If supplied, the result of a previous run with sigthresh=0 and totalpopest=F
...	control parameters such as mcmc, thin and seed passed to MCMCpoisson()

Details

This routine uses the routine MCMCpoisson() from the MCMCpack package. It works by casting the MSE problem into the form required by that routine and passing the MCMC parameters to that routine. It calls the routine cleanuplists.

bernardsilverman/modslavmse documentation built on Aug. 10, 2019, 2:39 a.m.