org.orangepalantir.leastsquares.fitters

Class LinearFitter

java.lang.Object

org.orangepalantir.leastsquares.fitters.LinearFitter

All Implemented Interfaces:

Fitter

public class LinearFitter
extends Object
implements Fitter

for solving linear least squares fit of f(x:A) = z with sets of x,z data points. derivatives are evaluated numerically and the function is assumed to be linear in A. Then a linear least squares fit is performed.

Constructor Summary

Constructors

Constructor and Description
LinearFitter(Function funct)

Method Summary

Modifier and Type	Method and Description
double	calculateDerivative(int k, double[] x) Given a set of parameters, and inputs, calculates the derivative of the k'th parameter d/d a_k (F) The function is linear in parameters eg: f(x) = params[0]* f0(x) + params[1]*f1(x) + ...
void	calculateDerivatives() Creates an array of derivatives since each one is used 3x's
double	calculateErrors() returns the cumulative error for current values
void	createAlphaMatrix()
void	createBetaMatrix()
void	fitData() Main routine, call this and the Parameters are iterated one time because the equations are assumed to be linear in parameter space.
double[][]	getCovarianceMatrix()
double[]	getParameters() Gets the current set of parameters values.
double[]	getUncertainty() Get an uncertainty value, similar to the way gnuplot does it.
void	initializeWorkspace()
void	iterateValues() Takes the current error, and the current parameter set and calculates the changes, then returns the maximum changed value
void	printMatrix()
void	setData(double[][] xvalues, double[] zvalues) Sets the values of of the original data points that are going to be fit.
void	setParameters(double[] parameters)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

LinearFitter

public LinearFitter(Function funct)

Method Detail

setData

public void setData(double[][] xvalues,
                    double[] zvalues)

Sets the values of of the original data points that are going to be fit.

Specified by:

setData in interface Fitter

setParameters

public void setParameters(double[] parameters)

Specified by:

setParameters in interface Fitter

calculateErrors

public double calculateErrors()

returns the cumulative error for current values

Specified by:

calculateErrors in interface Fitter

calculateDerivative

public double calculateDerivative(int k,
                                  double[] x)

Given a set of parameters, and inputs, calculates the derivative of the k'th parameter d/d a_k (F) The function is linear in parameters eg: f(x) = params[0]* f0(x) + params[1]*f1(x) + ... So the derivative k is fk(x) or the function evaluated when all of the parameters are zero except for the k'th parameter which is 1.

Parameters:

k - - index of the parameter that the derivative is being taken of

x - - set of values to use.

Returns:

derivative of function

calculateDerivatives

public void calculateDerivatives()

Creates an array of derivatives since each one is used 3x's

createBetaMatrix

public void createBetaMatrix()

createAlphaMatrix

public void createAlphaMatrix()

iterateValues

public void iterateValues()

Takes the current error, and the current parameter set and calculates the changes, then returns the maximum changed value

printMatrix

public void printMatrix()

initializeWorkspace

public void initializeWorkspace()

fitData

public void fitData()

Main routine, call this and the Parameters are iterated one time because the equations are assumed to be linear in parameter space.

Specified by:

fitData in interface Fitter

getParameters

public double[] getParameters()

Gets the current set of parameters values.

Specified by:

getParameters in interface Fitter

getUncertainty

public double[] getUncertainty()

Get an uncertainty value, similar to the way gnuplot does it. Essentially the diagonals of the covariance matrix.

Specified by:

getUncertainty in interface Fitter

Returns:

array of length of the parameters.

getCovarianceMatrix

public double[][] getCovarianceMatrix()