rsm.distributions: RSM Family Support Object
In marg: Approximate Marginal Inference for Regression-Scale Models

rsm.distributions

R Documentation

RSM Family Support Object

Description

Support object for regression-scale models

Details

This is a matrix object used in the construction of familyRsm objects. It is of mode list, and has dimension c(3,5). The 5 columns in

rsm.distributions[[functionName, familyName]]

represent the distributions student (Student's t), logistic, logWeibull, extreme (Gumbel or extreme value) and Huber (Huber's least favourable). The rows g0, g1 and g2 contain functions that define minus the corresponding log density and its first and second derivatives. They take a single argument for the families logistic, logWeibull and extreme, while for the families student and Huber they have a further argument, respectively df and k, which denote the degrees of freedom and the tuning constant.

This object is used by the functions make.family.rsm and, indirectly, by rsm when constructing a familyRsm object. See the documentation of familyRsm.object and make.family.rsm for additional details.