SmoothingSpline Class Reference

Interface to GCVSPL sub-routine by Herman J. Woltring. More...

#include <SmoothingSpline.h>

Public Member Functions
	SmoothingSpline ()
	Default constructor.
void	set_smoothing (double)
	Set the smoothing factor.
void	set_effective_nfree (double)
	Set the effective number of freely estimated parameters. More...
void	set_msre (double)
	Set the mean squared residual error, MSRE (Equation 5) More...
void	set_minimize_gcv (bool)
	Determine the smoothing factor using generalized cross-validation.
void	fit (const std::vector< double > &data_x, const std::vector< Estimate< double > > &data_y)
	Fit spline to data using current configuration.
unsigned	get_ndat_good () const
	Return the number of good data points included in the fit.
double	get_fit_gcv ()
double	get_fit_msre ()
double	get_fit_effective_nfree ()
double	get_fit_smoothing ()
double	get_fit_true_mse ()
double	get_fit_Gauss_Markov_error_variance ()
double	evaluate (double)
	evaluate the spline at the specified argument

Detailed Description

Interface to GCVSPL sub-routine by Herman J. Woltring.

Reference: H.J. Woltring (1986), A Fortran package for generalized, cross-validatory spline smoothing and differentiation. Adv. Eng. Software 8(3), 104-113

Member Function Documentation

set_effective_nfree()

void SmoothingSpline::set_effective_nfree

(

double

nfree

)

Set the effective number of freely estimated parameters.

Set the effective number of freely esimated parameters.

set_msre()

void SmoothingSpline::set_msre

(

double

msre

)

Set the mean squared residual error, MSRE (Equation 5)

Each term in equation 5 is multiplied by a weight, which is usually the inverse of the variance. In this case, Equation 5 is equivalent to the reduced chi-squared.

When the MSRE is specified, the smoothing factor is determined by minimizing the true predicted mean-squared error (Equation 6)

---

The documentation for this class was generated from the following files:

• SmoothingSpline.h
• SmoothingSpline.C