pspline.h File Reference

Smoothing using penalized Spline(P-Spline). More...

#include <stdexcept>
#include <vector>
#include "../misc/bspline.h"

Go to the source code of this file.

Functions
const std::vector< double >	sablib::PSpline (const std::vector< double > &y, const unsigned int knots_num, const unsigned int degree=3, const unsigned int s=2, const double lambda=1.0)
	Smoothes the input data using P-Splines (Penalized B-Splines).

Detailed Description

Smoothing using penalized Spline(P-Spline).

Author

Izadori

References:

• Eilers, P. H. C.; Marx, B. D. Flexible Smoothing with B-Splines and Penalties. Stastical Science, 1996, 11(2), 89-121.

Definition in file pspline.h.

Function Documentation

PSpline()

const std::vector< double > sablib::PSpline	(	const std::vector< double > &	y,
		const unsigned int	knots_num,
		const unsigned int	degree = 3,
		const unsigned int	s = 2,
		const double	lambda = 1.0 )

Smoothes the input data using P-Splines (Penalized B-Splines).

This function applies a B-Spline basis with a difference penalty on the coefficients to achieve smoothing of the input data.

Parameters

y	The input data points to be smoothed.
knots_num	The number of internal knots for the B-Spline basis.
degree	The degree of the B-Spline basis (default is 3, cubic spline).
s	The order of the difference penalty (default is 2).
lambda	The smoothing parameter (default is 1.0). Larger values result in smoother curves.

Returns

A vector containing the smoothed data points.

Exceptions

std::invalid_argument

One or more parameters are wrong.

Definition at line 15 of file pspline.cpp.

References sablib::BSpline< Scalar >::DesignMatrix(), and sablib::BSpline< Scalar >::Interpolate().