getPostStats: getPostStats

In hferg/bayestraitr: A set of tools to help with the interpretation and analysis of the output of BayesTraits MCMC analyses.

Description

Given a posterior of MCMC samples from a BayesTraits analysis (a logfile - typically with the .log.txt extension) and the name of one or morea parameters present in that logfile this will return the mean, median, mode and standard deviation of the parameter estimates.

Usage

getPostStats(logfile, params, thinning = 1, burnin = 0)

Arguments

logfile	The name of the logfile of the BayesTraits analysis.
params	The name(s) of one or more parameters present in the logfile.
thinning	Thinning parameter for the posterior - defaults to 1 (all samples). 2 uses every second sample, 3 every third and so on.
burnin	The number of generations to remove from the start of the chain as burnin. Use if the chain has not reached convergence before sampling began. Useful if the burnin parameter for the analysis itself was not long enough.

Value

An object of class data.frame with a row for each parameter, and columns for the mean, median, mode and standard deviation of those parameters.

Examples

params <- getParams("cool-data.log.txt")
getPostMean("cool-data.log.txt", params[c(2:5)]
getPostMean("cool-data.log.txt", "Lh")

hferg/bayestraitr documentation built on May 28, 2019, 8:55 p.m.