Skip to contents

The WSDM Function for GLMs

wsdm3.Rd

Computes the WSDM statistic for one or more parameters in a fitted generalized linear model. The object has "glm". The functions comprise an S3 generic and method.

Usage

wsdm3(object, ...)
# Default S3 method
wsdm3(object, ...)
# S3 method for class 'glm'
wsdm3(object, hdiff = 0.005,
     retry = TRUE, mux.hdiff = 1,
     maxderiv = 5, theta0 = 0, use.hdeff = FALSE,
     doffset = NULL, subset = NULL,
     derivs.out = FALSE, fixed.hdiff = TRUE,
     eps.wsdm = 0.15, Mux.div = 3, warn.retry = TRUE,
     with1 = TRUE, ...)

Arguments

object

a fitted GLM, created by glm.

hdiff,retry,mux.hdiff

Same as wsdm.

maxderiv,theta0

Same as wsdm.

use.hdeff,doffset

Same as wsdm.

subset,derivs.out

Same as wsdm.

fixed.hdiff

Same as wsdm.

eps.wsdm,Mux.div

Same as wsdm.

warn.retry, with1

Same as wsdm.

...

additional argument(s) for methods.

Value

See wsdm.

Details

While VGAM is written in S4, this wsdm3 is an S3 generic function so that the WSDM statistics can be computed for glm models. Since vglm and glm have different convergence criteria, computing the WSDM statistics for effectively the same model may differ.

See also

cops, coef, wsdm, fbeetle.

Examples

if (FALSE) { # \dontrun{
fit1 <- glm(cbind(dead, n-dead) ~ logdose, binomial, fbeetle, tr = TRUE)
coef(fit1)
wsdm3(fit1)
fit2 <- vglm(cbind(dead, n-dead) ~ logdose, binomialff, fbeetle, tr = TRUE)
coef(fit2)
wsdm(fit2)
} # }