This is a polygonal fitter for any number of in- and output dimensions allowing for inter- and extrapolation. More...

#include <PolyFit.hh>

Classes
struct	Impl

Public Types
typedef Eigen::Matrix< double, Eigen::Dynamic, Eigen::Dynamic >	polco
	Matrix holding the polynomial coefficients, one row for each output dimension.

Public Member Functions
	PolyFit (size_t idim, size_t odim, size_t ord)
	Creates a PolyFit instance for the specified number of input and output dimensions, and polynomial order.

	PolyFit (const std::vector< std::vector< double > > &invals, const std::vector< std::vector< double > > &outvals, size_t ord)
	Creates a PolyFit instance and adds the specified sample values.

void	addSample (const std::vector< double > &inval, const std::vector< double > &outval)
	Add a sample specifying the relation between input and output values.

polco	estimateCoefficients () const
	Estimate and return the coefficients of the polynomial that best fits the data.

std::vector< double >	getValue (const std::vector< double > &inval) const
	Estimates the polynomial result from the specified input.

Eigen::MatrixXd	getJacobian (const std::vector< double > &in_values) const
	Estimates the polynomial Jacobian from the specified input.

double	getSingle (double inval) const
	Convenience method for 1D polynomials.

double	getDerivative (double inval) const
	Estimates the polynomial derivative from the specified input for 1D polynomials.

void	clear ()
	Resets the contents of this object.

Detailed Description

This is a polygonal fitter for any number of in- and output dimensions allowing for inter- and extrapolation.

Member Typedef Documentation

◆ polco

typedef Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> gramods::gmMisc::PolyFit::polco

Matrix holding the polynomial coefficients, one row for each output dimension.

Constructor & Destructor Documentation

◆ PolyFit() [1/2]

gramods::gmMisc::PolyFit::PolyFit	(	size_t	idim,
		size_t	odim,
		size_t	ord
	)

Creates a PolyFit instance for the specified number of input and output dimensions, and polynomial order.

Parameters

[in]	idim	Input dimensionality.
[in]	odim	Output dimensionality.
[in]	ord	Order of the polynomial.

◆ PolyFit() [2/2]

gramods::gmMisc::PolyFit::PolyFit	(	const std::vector< std::vector< double > > &	invals,
		const std::vector< std::vector< double > > &	outvals,
		size_t	ord
	)

Creates a PolyFit instance and adds the specified sample values.

Member Function Documentation

◆ addSample()

void gramods::gmMisc::PolyFit::addSample	(	const std::vector< double > &	inval,
		const std::vector< double > &	outval
	)

Add a sample specifying the relation between input and output values.

The dimensions must match the dimensionality of the object or exception will be thrown.

Parameters

[in]	inval	The parameter value that should give rise to the specified output value.
[in]	outval	The output value for the specified inval.

◆ clear()

void gramods::gmMisc::PolyFit::clear ( )

Resets the contents of this object.

◆ estimateCoefficients()

PolyFit::polco gramods::gmMisc::PolyFit::estimateCoefficients ( ) const

Estimate and return the coefficients of the polynomial that best fits the data.

The order for these components matches the recursive expansion of [ 1, x, x², ... ] × [ 1, y, y², ... ] × ... into [ 1, x, x², ..., yx, yx², ..., y²x, y²x², ... ] (column wise) for each output (row wise).

This method will cache the coefficients until new samples have been added.

◆ getDerivative()

double gramods::gmMisc::PolyFit::getDerivative ( double inval ) const

Estimates the polynomial derivative from the specified input for 1D polynomials.

This method calls estimateCoefficients (if necessary) and calculates the resulting polynomial with the specified input.

◆ getJacobian()

Eigen::MatrixXd gramods::gmMisc::PolyFit::getJacobian ( const std::vector< double > & in_values ) const

Estimates the polynomial Jacobian from the specified input.

This method calls estimateCoefficients (if necessary) and calculates the Jacobian of the polynomial with the specified input. For n input variables, x₀ to xₙ, and m output variables, y₀ to yₘ, the returned matrix follows the convention

$\left( \begin{array}{cccc} \frac{\partial y_0}{\partial x_0} & \frac{\partial y_0}{\partial x_1} & \dots & \frac{\partial y_0}{\partial x_n} \\ \frac{\partial y_1}{\partial x_0} & \frac{\partial y_1}{\partial x_1} & \dots & \frac{\partial y_1}{\partial x_n} \\ \vdots \\ \frac{\partial y_m}{\partial x_0} & \frac{\partial y_m}{\partial x_1} & \dots & \frac{\partial y_m}{\partial x_n} \end{array} \right)$

The dimensions of the input must match the dimensionality of the object or exception will be thrown.

◆ getSingle()

double gramods::gmMisc::PolyFit::getSingle ( double inval ) const

Convenience method for 1D polynomials.

See also: getValue(const std::vector<double>&)

◆ getValue()

std::vector< double > gramods::gmMisc::PolyFit::getValue ( const std::vector< double > & inval ) const

Estimates the polynomial result from the specified input.

This method calls estimateCoefficients (if necessary) and calculates the resulting polynomial with the specified input. This method can be used both for inter- and extrapolation.