Wald confidence intervals

Wald intervals \(\hat\theta_j \pm z_{1 - \alpha/2} SE_j\) on the link scale, with \(SE_j\) from vcov.brs (Lopes, 2023, "Inferencia").

Usage

# S3 method for class 'brs'
confint(
  object,
  parm,
  level = 0.95,
  model = c("full", "mean", "precision"),
  ...
)

Arguments

object: A fitted "brs" object.
parm: Character or integer: which parameters. If missing, all parameters of model are returned.
level: Confidence level (default 0.95).
model: Character: "full", "mean" or "precision".
...: Currently ignored.

Value

Matrix with the lower and upper limits.

Details

Intervals for a mean or precision on the response scale follow by the inverse link of the limits (monotone links). A limit is NA when the variance is not estimable (see 'Fit diagnostics' in brs). For small samples or parameters near the border of the scale, brs_bootstrap gives intervals that do not rely on the normal approximation.

References

Lopes, J. E. (2023). Modelos de regressao beta para dados de escala. Master's dissertation, Universidade Federal do Parana, Curitiba. URI: https://hdl.handle.net/1884/86624.

Examples

set.seed(2023)
d <- data.frame(x = runif(150))
s <- brs_sim(~ x, data = d, beta = c(-0.5, 1), phi = qlogis(0.3), ncuts = 10)
fit <- brs(y ~ x, data = s)
confint(fit)
#>                  2.5 %     97.5 %
#> (Intercept) -0.5662544  0.1735143
#> x           -0.2500579  0.9946682
#> (phi)       -1.0194170 -0.5981087
# Mean at x = 0 on (0, 1): inverse logit of the intercept limits
plogis(confint(fit, parm = "(Intercept)"))
#>                 2.5 %    97.5 %
#> (Intercept) 0.3621016 0.5432701