rstar.ci.control {likelihoodAsy}

Auxiliary function for controlling computation of r*-based confidence intervals

Description

Auxiliary function for rstar.ci.

Usage

rstar.ci.control(npoints=10, away=0.3, stepsizefac=3, maxstep=50, trace=TRUE)

Arguments

npoints

Integer giving the number of points at which the r* and r statistics will be evaluated away from the MLE in each direction when both lower and upper are not null. When either lower or upper are NULL, such value is only roughly proportional to the number of evaluation points. Default is 10.

away

Positive value indicating how far from the MLE the computation of the r* and r statistics will be started, expressed in units of standard error of the scalar function of interest. Default is 0.3.

stepsizefac

Positive value used to determine the stepsize of the confidence interval algorithm when either lower or upper are null. In particular, the stepsize is given by stepsizefac/npoints times the standard error of the scalar function of interest. Default is 3.

maxstep

Integer giving the maximum number of steps that will be taken for crossing the normal quantiles for a 99 confidence interval for the r* statistic. Default is 50.

trace

Logical indicating if output should be produced during the computation. Default is TRUE.

Value

A list with components named as the arguments.

See Also

rstar.ci.

Examples

# A variation on example(rstar.ci):
## Not run: 
obj <- rstar.ci(quinedata, thetainit=c(coef(quine.nb1),quine.nb1$theta),  floglik=logLikNbin, 
                datagen=genDataNbin, fscore=gradLikNbin, fpsi=function(theta) theta[2], R=1000, 
                psidesc="Coefficient of EthN", npoints=5, away=0.1)
plot(obj)

## End(Not run)