mdyplControl: Auxiliary function for 'glm()' fitting using the 'brglmFit()'...

In brglm2: Bias Reduction in Generalized Linear Models

Auxiliary function for glm() fitting using the brglmFit() method.

Description

Typically only used internally by brglmFit(), but may be used to construct a control argument.

Usage

mdyplControl(alpha = NULL, epsilon = 1e-08, maxit = 25, trace = FALSE)

mdypl_control(alpha = NULL, epsilon = 1e-08, maxit = 25, trace = FALSE)

Arguments

alpha	the shrinkage parameter (in ⁠[0, 1]⁠) in the Diaconis-Ylvisaker prior penalty. Default is NULL, which results in alpha = m / (m + p), where m is the sum of the binomial totals and p is the number of model parameters. Setting alpha = 1 corresponds to using maximum likelihood, i.e. no penalization. See Details.
epsilon	positive convergence tolerance epsilon. Default is 1e-08.
maxit	integer giving the maximal number of iterations allowed. Default is 25.
trace	logical indicating if output should be produced for each iteration. Default is FALSE.

Details

Internally, mdyplFit() uses stats::glm.fit() to fit a logistic regression model on responses alpha * y + (1 - alpha) / 2, where y are the original binomial responses scaled by the binomial totals. epsilon, maxit and trace control the stats::glm.fit() call; see stats::glm.control().

Value

A list with components named as the arguments.

Author(s)

Ioannis Kosmidis ⁠[aut, cre]⁠ [e-mail]

See Also

mdyplFit(), glm.control()

brglm2 documentation built on Aug. 29, 2025, 5:25 p.m.