confint: Confidence intervals and profile likelihoods for parameters...

In perbrock/sensR: Thurstonian Models for Sensory Discrimination

Confidence intervals and profile likelihoods for parameters in 2AC models

Description

Computes confidence intervals from the profiled likelihood and the Wald approximation in the 2AC model, or plots the profile likelihood function for d.prime.

Usage

## S3 method for class 'twoAC'
confint(object, parm, level = 0.95,
        type = c("likelihood", "Wald"), ...)

## S3 method for class 'profile.twoAC'
confint(object, parm = "d.prime", level = 0.95, ...)

## S3 method for class 'twoAC'
profile(fitted, alpha = 1e-3, nSteps = 1e2, range, ...)

## S3 method for class 'profile.twoAC'
plot(x, level = c(0.95, 0.99), Log = FALSE,
      relative = TRUE, fig = TRUE, n = 1e3, ..., ylim = NULL)

Arguments

object	a fitted twoAC object or a profile.twoAC object.
fitted	a fitted twoAC object.
x	a profile.twoAC object.
type	the type of confidence interval required. "profile" is the most accurate.
parm	For confint.profile.twoAC: has to be "d.prime". For confint.twoAC: for type = "Wald" a specification of which parameters the confidence interval is required for. Ignored for type = "profile".
level	the confidence level required.
alpha	determines the range of profiling. By default the likelihood is profiled in the 99.9% Wald confidence interval region.
range	if supplied, d.prime will be profiled between min(range) and max(range). This over-rules the automatic range computation.
nSteps	the number of profile steps.
Log	should the profile likelihood be plotted on the log-scale?
relative	should the relative or the absolute likelihood be plotted?
fig	should the profile likelihood be plotted?
n	the no. points used in the spline interpolation of the profile likelihood.
ylim	overrules default y-limits on the plot of the profile likelihood.
...	not currently used.

Details

These confint methods call the appropriate profile method, then finds the confidence intervals by interpolation of the profile traces. If the profile object is already available, this should be used as the main argument rather than the fitted model object itself.

In plot.profile.twoAC: at least one of Log and relative arguments have to be TRUE.

Value

confint: A matrix (or vector) with columns giving lower and upper confidence limits for each parameter. These will be labelled as (1-level)/2 and 1 - (1-level)/2 in % (by default 2.5% and 97.5%). Profile likelihood confindence intervals are only available for d.prime and not tau.

profile.twoAC: a data.frame with the profile of d.prime.

plot.profile.twoAC invisibly returns the spline approcimation to the profile.

Author(s)

Rune Haubo B Christensen

References

Christensen R.H.B., lee H-S and Brockhoff P.B. (2012). Estimation of the Thurstonian model for the 2-AC protocol. Food Quality and Preference, 24(1), pp.119-128.

See Also

profile and confint

Examples

(fm1 <- twoAC(c(2, 2, 6)))
confint(fm1)
confint(fm1, type = "Wald")

pr1 <- profile(fm1)
confint(pr1)

pr1 <- profile(fm1, alpha = 1e-5)
old <- par(mfrow = c(2,2))
plot(pr1)
plot(pr1, Log = FALSE, relative = TRUE)
plot(pr1, Log = TRUE, relative = TRUE)
plot(pr1, Log = TRUE, relative = FALSE)
par(old)

perbrock/sensR documentation built on Nov. 5, 2023, 10:41 a.m.