computeESS: Compute the ESS for the MCMC samples

In Caetanods/ratematrix: Bayesian Estimation of the Evolutionary Rate Matrix

Compute the ESS for the MCMC samples

Description

Computes the Effective Sample Size (ESS) for the parameters of the model from the MCMC samples.

Usage

computeESS(mcmc, p)

Arguments

mcmc	Posterior distribution object. Same as output from 'readMCMC' function.
p	Number of evolutionary rate matrix regimes fitted to the phylogenetic tree.

Details

Function uses 'coda' function 'effectiveSize' to compute the ESS for each of the parameters of the model separatelly. Values for the ESS is too low indicates poor mixing for the parameter of the model.

Value

A list object with the ESS value for the root, evolutionary rates, and evolutionary correlations among the traits.

Author(s)

Daniel Caetano and Luke Harmon

Examples

library( ratematrix )
data( centrarchidae )
## Run multiple MCMC chains.
handle.list <- lapply(1:4, function(x) ratematrixMCMC(data=centrarchidae$data
                      , phy=centrarchidae$phy.map, gen=10000, dir=tempdir()) )
## Read all to a list
posterior.list <- lapply(handle.list, function(x) readMCMC(handle = x, burn = 0.5))
## Merge all posteriors.
merged.four <- mergePosterior(posterior.list)
## Compute the ESS for the merged posterior:
ess <- computeESS(mcmc=merged.four, p=2)

Caetanods/ratematrix documentation built on Jan. 4, 2023, 3:12 p.m.