Method for Profiling vglm Objects — profilevglm • VGAM

Investigates the profile log-likelihood function for a fitted model of class "vglm".

Usage

profilevglm(object, which = 1:p.vlm, alpha = 0.01,
            maxsteps = 10, del = zmax/5, trace = NULL, ...)

Arguments

object

the original fitted model object.

which

the original model parameters which should be profiled. This can be a numeric or character vector. By default, all parameters are profiled.

alpha

highest significance level allowed for the profiling.

maxsteps

maximum number of points to be used for profiling each parameter.

del

suggested change on the scale of the profile t-statistics. Default value chosen to allow profiling at about 10 parameter values.

trace

logical: should the progress of profiling be reported? The default is to use the trace value from the fitted object; see vglm.control for details.

...

further arguments passed to or from other methods.

Value

A list of classes "profile.glm" and "profile" with an element for each parameter being profiled. The elements are data-frames with two variables

par.vals: a matrix of parameter values for each fitted model.
tau: the profile t-statistics.

Details

This function is called by confintvglm to do the profiling. See also profile.glm for details.

Author

T. W. Yee adapted this function from profile.glm, written originally by D. M. Bates and W. N. Venables. (For S in 1996.) The help file was also used as a template.

Examples

pneumo <- transform(pneumo, let = log(exposure.time))
fit1 <- vglm(cbind(normal, mild, severe) ~ let, propodds,
             trace = TRUE, data = pneumo)
#> Iteration 1: deviance = 5.10322
#> Iteration 2: deviance = 5.026838
#> Iteration 3: deviance = 5.026826
#> Iteration 4: deviance = 5.026826
pfit1 <- profile(fit1, trace = FALSE)
confint(fit1, method = "profile", trace = FALSE)
#>                    2.5 %    97.5 %
#> (Intercept):1 -12.491514 -7.300884
#> (Intercept):2 -13.436780 -8.165407
#> let             1.907272  3.401708