calculateGelman: Calculate the Gelman and Rubin diagnostic for bandle output

In ococrook/bandle: An R package for the Bayesian analysis of differential subcellular localisation experiments

Calculate the Gelman and Rubin diagnostic for bandle output

Description

This function is a wrapper function for the gelman.diag function from the coda package. It takes a bandleParams object and calculates the Gelman and Rubin's convergence diagnostic (otherwise known as the potential scale reduction factor) for all pairwise MCMC chain combinations, together with upper and lower confidence limits.

Usage

calculateGelman(params)

Arguments

params

An instance of class bandleParams

Value

A list of 2 matrix array's, one for each condition containing the point estimates of the potential scale reduction factor (labelled Point est.) and their upper confidence limits (labelled Upper C.I.).

Examples

## Generate some example data
library("pRolocdata")
data("tan2009r1")
set.seed(1)
tansim <- sim_dynamic(object = tan2009r1,
                      numRep = 4L,
                      numDyn = 100L)
data <- tansim$lopitrep
control <- data[1:2]
treatment <- data[3:4]

## fit GP params
gpParams <- lapply(tansim$lopitrep, function(x)
fitGPmaternPC(x, hyppar = matrix(c(0.5, 1, 100), nrow = 1)))

## run bandle
res <- bandle(objectCond1 = control,
              objectCond2 = treatment,
              gpParams = gpParams,
              fcol = "markers",
              numIter = 20L,
              burnin = 1L,
              thin = 2L,
              numChains = 2,
              BPPARAM = SerialParam(RNGseed = 1),
              seed = 1)

## Process the results
calculateGelman(res)

ococrook/bandle documentation built on Nov. 4, 2024, 12:27 a.m.