ptycho (version 1.1-4)

checkConvergence: Compute Differences Between MCMC Chains

Description

Compute the differences between the chains in a ptycho object of the means of $\tau$ and the indicator variables.

Usage

checkConvergence(obj, doLastIterOnly=TRUE)

Arguments

obj

A ptycho object

doLastIterOnly

Logical specifying whether to compute differences only for the last MCMC iteration in the input object or for all iterations

Value

A data frame with the following columns:

iter

MCMC iteration number

type

Factor specifying the type of the variable; one of “tau”, “var” for variant indicator variable, or “grp” for second-level indicator variable

index

Number specifying the pertinent column in the design matrix (for type equal to “var” or for type equal to “grp” when Across Traits prior was used) or the variant group index (for type equal to “grp” when Across Sites prior was used); equal to 1 for type equal to “tau”

y

Factor specifying the name of the response; empty for type equal to “tau” or for type equal to “grp” when Across Traits prior was used

range

Difference between maximum and minimum across chains

Details

For $\tau$ and each indicator variable in the input ptycho object, compute the difference between the maximum and the minimum mean for each chain. If doLastIterOnly is TRUE, then the differences are only computed for the last iteration in each chain; otherwise, the differences are computed at each iteration in the input object.

See Also

ptycho; also ptychoOut for example below

Examples

data(ptychoOut)
cvg <- checkConvergence(ptychoOut, doLastIterOnly=FALSE)
reshape2::dcast(cvg, ... ~ iter, value.var="range")