tmbroot: Compute likelihood profile confidence intervals of a TMB...

In TMB: Template Model Builder: A General Random Effect Tool Inspired by 'ADMB'

Compute likelihood profile confidence intervals of a TMB object by root-finding

Description

Compute likelihood profile confidence intervals of a TMB object by root-finding in contrast to tmbprofile, which tries to compute somewhat equally spaced values along the likelihood profile (which is useful for visualizing the shape of the likelihood surface), and then (via confint.tmbprofile) extracting a critical value by linear interpolation,

Usage

tmbroot(
  obj,
  name,
  target = 0.5 * qchisq(0.95, df = 1),
  lincomb,
  parm.range = c(NA, NA),
  sd.range = 7,
  trace = FALSE,
  continuation = FALSE
)

Arguments

obj	Object from MakeADFun that has been optimized.
name	Name or index of a parameter to profile.
target	desired deviation from minimum log-likelihood. Default is set to retrieve the 95 if the objective function is a negative log-likelihood function
lincomb	Optional linear combination of parameters to profile. By default a unit vector corresponding to name.
parm.range	lower and upper limits; if NA, a value will be guessed based on the parameter value and sd.range
sd.range	in the absence of explicit parm.range values, the range chosen will be the parameter value plus or minus sd.range times the corresponding standard deviation. May be specified as a two-element vector for different ranges below and above the parameter value.
trace	report information?
continuation	use continuation method, i.e. set starting parameters for non-focal parameters to solutions from previous fits?

Value

a two-element numeric vector containing the lower and upper limits (or NA if the target is not achieved in the range), with an attribute giving the total number of function iterations used

Examples

## Not run: 
runExample("simple",thisR=TRUE)
logsd0.ci <- tmbroot(obj,"logsd0")

## End(Not run)

TMB documentation built on Sept. 11, 2024, 7:06 p.m.