Set control parameters for cumulative link models
clm.control(method = c("Newton", "model.frame", "design", "ucminf", "nlminb",
"optim"),
sign.location = c("negative", "positive"),
sign.nominal = c("positive", "negative"),
..., trace = 0L,
maxIter = 100L, gradTol = 1e-06, maxLineIter = 15L, relTol = 1e-6,
tol = sqrt(.Machine$double.eps), maxModIter = 5L,
convergence = c("warn", "silent", "stop", "message"))
a list of control parameters.
"Newton"
fits the model by maximum likelihood and
"model.frame"
cause clm
to return the
model.frame
, "design"
causes clm
to
return a list of design matrices etc. that can be used with
clm.fit
. ucminf
, nlminb
and optim
refer
to general purpose optimizers.
change sign of the location part of the model.
change sign of the nominal part of the model.
numerical, if > 0
information is printed about and during
the optimization process. Defaults to 0
.
the maximum number of Newton-Raphson iterations.
Defaults to 100
.
the maximum absolute gradient; defaults to 1e-6
.
the maximum number of step halfings allowed if
a Newton(-Raphson) step over shoots. Defaults to 15
.
relative convergence tolerence: relative change in the
parameter estimates between Newton iterations. Defaults to 1e-6
.
numerical tolerence on eigenvalues to determine negative-definiteness of Hessian. If the Hessian of a model fit is negative definite, the fitting algorithm did not converge. If the Hessian is singular, the fitting algorithm did converge albeit not to a unique optimum, so one or more parameters are not uniquely determined even though the log-likelihood value is.
the maximum allowable number of consecutive
iterations where the Newton step needs to be modified to be a decent
direction. Defaults to 5
.
action to take if the fitting algorithm did not converge.
Rune Haubo B Christensen
clm