runLossDevModel,LossDevModelInput-method {BALD}R Documentation

A method to run models in BALD.

Description

A method to run models in BALD.

Arguments

object

The object of type LossDevModelInput containing the model to estimate.

burnIn

An integer to represent the number of initial MCMC iterations to be discarded. (The adaptive phase (nAddapt) is not considered part of burnIn.)

sampleSize

An integer to represent the number of MCMC iterations to execute following the burnIn. (Actual number of iterations kept approximately sampleSize / thin.)

thin

Keep every thin'th value of sampleSize.

nChains

The number of MCMC chains to run.

nAddapt

The length of the adaptive phase for the MCMC algorithm. (Default is trunc(burnIn/4)+1.)

Details

This method returns a valid output object or flags an error. This method is suitable for classes properly derived from class LossDevModelInput that properly override “getJagsData” and “getJagsInits” and whose output type has a valid getModelOutputNodes method.

BALD sets the seed in each chain from a random number generated inside of R. So to make a run reproducible, all one must do is set the seed (using set.seed) in R prior to the execution of this method.

Value

An object of class LossDevModelOutput.

See Also

runLossDevModel set.seed


[Package BALD version 1.0.0-3 Index]