wald_univariate: Univariate Wald Tests and Confidence Intervals for lgspline...
In lgspline: Lagrangian Multiplier Smoothing Splines for Smooth Function Estimation

wald_univariate

R Documentation

Univariate Wald Tests and Confidence Intervals for lgspline Coefficients

Description

Computes per-coefficient Wald tests and confidence intervals from a fitted lgspline. For Gaussian identity-link models, t-statistics and t-intervals are used; otherwise z-statistics.

Usage

wald_univariate(object, scale_vcovmat_by = 1, cv, ...)

Arguments

`object`	A fitted lgspline object. Must have been fit with `return_varcovmat = TRUE`.
`scale_vcovmat_by`	Numeric; scaling factor for the variance-covariance matrix. Default 1.
`cv`	Numeric; critical value for confidence intervals. If missing, defaults to `object$critical_value` or `qnorm(0.975)`.
`...`	Additional arguments passed to the internal `wald_univariate` method.

Value

An object of class "wald_lgspline", a list with:

coefficients: Matrix with columns: Estimate, Std. Error, t value or z value, Pr(>|t|) or Pr(>|z|), CI LB, CI UB.
critical_value: Critical value used.
family: GLM family from the fitted model.
N: Number of observations.
trace_XUGX: Effective df trace term.
statistic_name: "t value" or "z value".
p_value_name: "Pr(>|t|)" or "Pr(>|z|)".
df.residual: Residual degrees of freedom when supplied by the internal Wald method.

Print, summary, and plot methods are available; see print.wald_lgspline, summary.wald_lgspline, plot.wald_lgspline.

Examples


set.seed(1234)
t <- runif(1000, -10, 10)
y <- 2*sin(t) + -0.06*t^2 + rnorm(length(t))
model_fit <- lgspline(t, y, return_varcovmat = TRUE)

wald_default <- wald_univariate(model_fit)
print(wald_default)

## t-distribution critical value
eff_df <- model_fit$N - model_fit$trace_XUGX
wald_t <- wald_univariate(model_fit, cv = qt(0.975, eff_df))
print(wald_t)

coef_table <- wald_default$coefficients
plot(wald_default)

lgspline documentation built on May 8, 2026, 5:07 p.m.