Penalized Basis Smoothing — smooth.basis.fd • fdars

Smooth functional data using penalized B-spline or Fourier basis expansion. Penalized Basis Smoothing

Usage

smooth.basis.fd(
  fdataobj,
  type = c("bspline", "fourier"),
  nbasis = NULL,
  lambda = 0,
  lfd.order = 2,
  period = NULL
)

Arguments

fdataobj: An object of class 'fdata'.
type: Basis type: "bspline" or "fourier".
nbasis: Number of basis functions. If NULL, defaults to min(ncol/4, 30).
lambda: Smoothing parameter (penalty weight). Use 0 for no penalty.
lfd.order: Order of the linear differential operator for the penalty (default 2 = penalize curvature).
period: Period for Fourier basis (only used when type = "fourier"). If NULL, defaults to the range of argvals.

Value

An object of class 'smooth.basis' with components:

coefficients: Matrix of basis coefficients (nbasis x n)
fitted: Fitted smoothed curves as an fdata object
edf: Effective degrees of freedom
gcv: Generalized cross-validation score
aic: Akaike information criterion
bic: Bayesian information criterion
nbasis: Number of basis functions used
lambda: Smoothing parameter used
type: Basis type used
fdataobj: Original functional data

Details

Fits a penalized basis expansion to smooth functional data. Supports B-spline and Fourier basis types with a roughness penalty.

Examples

# \donttest{
fd <- fdata(matrix(rnorm(500), 50, 10), argvals = seq(0, 1, length.out = 10))
sm <- smooth.basis.fd(fd, type = "bspline", nbasis = 5, lambda = 0.1)
sm
#> Penalized Basis Smoothing
#>   Basis type: bspline 
#>   Number of basis functions: 6 
#>   Lambda: 0.1 
#>   EDF: 2.2 
#>   GCV: 1.37371 
# }