coef: Coef Method for GeDS Objects

In GeDS: Geometrically Designed Spline Regression

Coef Method for GeDS Objects

Description

Method for the function coef that allows to extract the estimated coefficients of a fitted GeDS regression model from a "GeDS" class object.

Usage

## S3 method for class 'GeDS'
coef(object, n = 3L, onlySpline = TRUE, ...)

Arguments

object	The "GeDS" class object from which the coefficients of the selected GeDS regression model should be extracted.
n	Integer value (2, 3 or 4) specifying the order (= degree + 1) of the "GeDS", "GeDSgam" or "GeDSboost" fit whose coefficients should be extracted. By default equal to 3L; non-integer values will be passed to the function as.integer.
onlySpline	Logical variable specifying whether only the coefficients for the GeDS component of a fitted multivariate regression model should be extracted or whether the coefficients of both the GeDS and the parametric components should be returned.
...	Potentially further arguments (required by the definition of the generic function). These will be ignored, but with a warning.

Details

Simple method for the function coef.

As "GeDS" class objects contain three different fits (linear, quadratic and cubic), the argument n can be used to specify the order of the GeDS fit for which regression coefficients are required.

As mentioned in the Details of formula, the predictor model may be multivariate and it may include a (univariate or bivariate) GeD spline component, plus a parametric component involving the remaining variables. If the onlySpline argument is set to TRUE (the default value), only the coefficients corresponding to the GeD spline component of order n of the multivariate predictor model are extracted.

Value

A named vector containing the required coefficients of the fitted univariate or multivariate predictor model. The coefficients corresponding to the variables that enter the parametric component of the fitted multivariate predictor model are named as the variables themselves. The coefficients of the GeDS component are coded as "N" followed by the index of the corresponding B-spline.

See Also

coef for the standard definition; NGeDS for more examples.

Examples

# Generate a data sample for the response variable
# and the covariates
set.seed(123)
N <- 500
f_1 <- function(x) (10*x/(1+100*x^2))*4+4
X <- sort(runif(N ,min = -2, max = 2))
Z <- runif(N)
# Specify a model for the mean of the response Y to be a superposition of
# a non-linear component f_1(X), a linear component 2*Z and a
# free term 1, i.e.
means <- f_1(X) + 2*Z + 1
# Add normal noise to the mean of Y
Y <- rnorm(N, means, sd = 0.1)

# Fit to this sample a predictor model of the form f(X) + Z, where
# f(X) is the GeDS component and Z is the linear (additive) component
# see ?formula.GeDS for details
(Gmod <- NGeDS(Y ~ f(X) + Z, beta = 0.6, phi = 0.995, Xextr = c(-2,2)))

# Extract the GeD spline regression coefficients
coef(Gmod, n = 3)

# Extract all the coefficients, including the one for the linear component
coef(Gmod, onlySpline = FALSE, n = 3)

GeDS documentation built on June 30, 2025, 9:07 a.m.