PENALTY-class {BSL}R Documentation

S4 class “PENALTY”

Description

This S4 class contains the penalty selection result from function selectPenalty. show display the penalty selection result. plot plot the penalty selection result using ggplot2.

Usage

## S4 method for signature 'PENALTY'
show(object)

## S4 method for signature 'PENALTY,ANY'
plot(x, logscale = TRUE)

## S4 method for signature 'BSL'
getPenalty(object)

Arguments

object

The S4 object of class “PENALTY” to show.

x

The S4 object of class “PENALTY” to plot.

logscale

A logical argument whether the x-axis (penalty) should be log transformed. The default is TRUE.

Slots

loglike

A list of the log-likelihood values. The list contains multiple matrices (each corresponds to the result for a specific n value). The number of row of the matrix equals to the number of repeats M. The columns of the matrix stands for different penalty values.

n

A vector of n, the number of simulations from the model per MCMC iteration for estimating the synthetic likelihood.

lambda

A list, with each entry containing the vector of penalty values for the corresponding choice of n.

M

The number of repeats used in estimating the standard deviation of the estimated log synthetic likelihood.

sigma

The standard deviation of the log synthetic likelihood estimator to aim for, usually a value between 1 and 2. This reflects the mixing of a Markov chain.

model

A “MODEL” object generated with function newModel. See newModel.

stdLoglike

A list contains the estimated standard deviations of log-likelihoods.

penalty

The vector stores the selected penalty values for each given n by choosing from the candidate lambda list. The selected values produce closest standard deviations stdLoglike to the target sigma.

result

The result data frame.

call

The original code used to run selectPenalty.

See Also

selectPenalty for the function that selects the penalty parameter.


[Package BSL version 3.2.4 Index]