lossDevOptions: Options for BALD.

In BALD: Robust Loss Development Using MCMC

Description

Options for BALD.

Usage

lossDevOptions(...)

Arguments

...	named values to set. If empty, only the current list of option settings is returned.

Details

Currently the only options are keepCodaOnDisk and logsplinePenaltyFunction.

logsplinePenaltyFunction

When drawing kernal density plots using the logspline, it maybe desirable to specify a penalty to smooth the density (See ?logspline). This value must be a function which takes one paramter (a vector of the sampled data points) and returns one value – the penalty. The default returns the the log of the number of draws.

Value

The current (or altered) list of option settings is returned.

Examples

library(BALD) 
#define the log of sample size function
logsamplesize <- function(x) {
log(length(x))
}
#assign the log of sample size function as penalty function
lossDevOptions(logsplinePenaltyFunction = logsamplesize)

BALD documentation built on May 2, 2019, 6:51 a.m.