profile.egf: Univariate Likelihood Profiles

Description

Computes univariate likelihood profiles of fixed effect coefficients, random effect covariance parameters, and linear combinations thereof, including population fitted values.

Usage

# S3 method for egf
profile(fitted, level = 0.95,
        A = seq_along(par), grid = 12L,
        parallel = egf_parallel(), trace = FALSE,
        top = egf_top(fitted), subset = NULL, select = NULL, ...)
# S3 method for profile.egf
confint(object, parm = seq_along(object), level = attr(object, "level"),
        class = FALSE, ...)
# S3 method for profile.egf
plot(x, parm = seq_along(x), level = attr(x, "level"),
     type = c("z", "abs(z)", "z^2"), ...)

Value

A list of length nrow(A) inheriting from classes profile.egf and profile. Each element is a data frame specifying a profile, with two variables:

z: a numeric vector containing profile \(z\)-statistics. The profile \(z\)-statistic is the appropriately signed square root of the change in deviance under the restricted model.
par.vals: a numeric matrix with one column containing values of the linear combination specified by A[i, ].

The confidence level level is preserved as an attribute.

Arguments

fitted: an egf object.
level: a number in the interval \((0,1)\) indicating a confidence level. Profiles are computed up to a change in deviance equal to qchisq(level, df = 1).
A: a numeric matrix with 1+p columns, where p = length(coef(fitted)), in which case each row specifies a linear combination of the elements of c(1, coef(fitted)) to be profiled; or a valid index vector for coef(fitted), in which case the indexed elements are profiled; or NULL, in which case population fitted values are profiled.
grid: a positive integer. Step sizes chosen adaptively by tmbprofile will generate approximately this many points on each side of a profile's minimum point.
parallel: an egf_parallel object defining options for R level parallelization.
trace: a logical. If TRUE, then basic tracing messages indicating progress are printed. These may be mixed with optimizer output depending on fitted[["control"]][["trace"]].
top: a subset of egf_top(fitted) naming top level nonlinear model parameters for which profiles on population fitted values should be profiled.
subset, select: index vectors for the rows and columns of model.frame(fitted, "combined") or language objects evaluating to such vectors. subset indicates fitting windows for which profiles should be computed; the default indicates all. select indicates variables that should be appended to the result; the default indicates none. Evaluation of language objects follows the implementation of subset.data.frame.
...: additional arguments passed from or to other methods.
object, x: a profile.egf object.
parm: a valid index vector for object or x indicating a subset of the profiles.
class: a logical. If TRUE and if object was created by profile(A = NULL), then the value of the method call is a confint.egf object, not a matrix.
type: a character string indicating which of \(z\), \(|z|\), and \(z^{2}\) is plotted.

Details

Computation of likelihood profiles is typically expensive, requiring estimation of many restricted models. It is parallelized at the C++ level when there is OpenMP support and fitted[["control"]][["omp_num_threads"]] is set to an integer greater than 1. When there is no OpenMP support, it can still be parallelized at the R level with appropriate setting of argument parallel.

Examples

Run this code

# \donttest{
example("egf", package = "epigrowthfit")

zz <- profile(m1, A = NULL,
              top = "log(r)", subset = quote(country == "A" & wave == 1))
str(zz)

confint(zz, class = TRUE)

pty <- c("z", "abs(z)", "z^2")
bty <- c("l", "u", "u")
for (i in 1:3)
    plot(zz, type = pty[i], bty = bty[i], las = 1)
# }