DiffBetweenEstimates: Difference between estimated parameters

In oswaldosantos/ptb: Prevalence Estimation and Diagnostic Test Evaluation in a Bayesian Framework

Description

Vectorized difference between chains of two objects of class mcmc.list.

Usage

DiffBetweenEstimates(estimates)

Arguments

estimates

list with two objects of class mcmc.list. These objects must have the same number of chains and the chains must have the same length.

Value

A list of class mcmc.list.

Examples

# Priors.
priors <- c(true_prev_a = 1, true_prev_b = 1,
            se_a = 6.28, se_b = 13.32, sp_a = 212.12, sp_b = 3.13)

# First estimate.
dataset1 <- list(pop_size = 100, positives = 5)
prev_est1 <- OneTestOnePopBM(dataset = as.list(dataset1), n_iter = 3e3,
                                  priors = priors, pars = "true_prev",
                                  burn_in = 5e2)

# Second estimate.
dataset2 <- list(pop_size = 91, positives = 1)
prev_est2 <- OneTestOnePopBM(dataset = as.list(dataset2), n_iter = 3e3,
                                   priors = priors, pars = "true_prev",
                                   burn_in = 5e2)

# Estimated difference.
diffs <- DiffBetweenEstimates(list(prev_est1, prev_est2))
summary(diffs)

# Diagnostic plots.
library(coda); library(ggmcmc)
gelman.diag(diffs)
gelman.plot(diffs)
gg_res <- ggs(diffs)
ggs_traceplot(gg_res)
ggs_density(gg_res)
ggs_histogram(gg_res, bins = 100)
ggs_compare_partial(gg_res)
ggs_running(gg_res)
ggs_autocorrelation(gg_res)

oswaldosantos/ptb documentation built on May 24, 2019, 5:13 p.m.