org.apache.commons.math.optimization.fitting

Class CurveFitter

java.lang.Object

org.apache.commons.math.optimization.fitting.CurveFitter

public class CurveFitter
extends java.lang.Object

Fitter for parametric univariate real functions y = f(x).

When a univariate real function y = f(x) does depend on some unknown parameters p₀, p₁ ... p_n-1, this class can be used to find these parameters. It does this by fitting the curve so it remains very close to a set of observed points (x₀, y₀), (x₁, y₁) ... (x_k-1, y_k-1). This fitting is done by finding the parameters values that minimizes the objective function ∑(y_i-f(x_i))². This is really a least squares problem.

Since:

2.0

Constructor Summary

Constructors

Constructor and Description
CurveFitter(DifferentiableMultivariateVectorialOptimizer optimizer) Simple constructor.

Method Summary

Modifier and Type	Method and Description
void	addObservedPoint(double x, double y) Add an observed (x,y) point to the sample with unit weight.
void	addObservedPoint(double weight, double x, double y) Add an observed weighted (x,y) point to the sample.
void	addObservedPoint(WeightedObservedPoint observed) Add an observed weighted (x,y) point to the sample.
void	clearObservations() Remove all observations.
double[]	fit(ParametricRealFunction f, double[] initialGuess) Fit a curve.
WeightedObservedPoint[]	getObservations() Get the observed points.

Methods inherited from class java.lang.Object

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

Constructor Detail

CurveFitter

public CurveFitter(DifferentiableMultivariateVectorialOptimizer optimizer)

Simple constructor.

Parameters:

optimizer - optimizer to use for the fitting

Method Detail

addObservedPoint

public void addObservedPoint(double x,
                             double y)

Add an observed (x,y) point to the sample with unit weight.

Calling this method is equivalent to call addObservedPoint(1.0, x, y).

Parameters:

x - abscissa of the point

y - observed value of the point at x, after fitting we should have f(x) as close as possible to this value

See Also:

addObservedPoint(double, double, double), addObservedPoint(WeightedObservedPoint), getObservations()

addObservedPoint

public void addObservedPoint(double weight,
                             double x,
                             double y)

Add an observed weighted (x,y) point to the sample.

Parameters:

weight - weight of the observed point in the fit

x - abscissa of the point

y - observed value of the point at x, after fitting we should have f(x) as close as possible to this value

See Also:

addObservedPoint(double, double), addObservedPoint(WeightedObservedPoint), getObservations()

addObservedPoint

public void addObservedPoint(WeightedObservedPoint observed)

Add an observed weighted (x,y) point to the sample.

Parameters:

observed - observed point to add

See Also:

addObservedPoint(double, double), addObservedPoint(double, double, double), getObservations()

getObservations

public WeightedObservedPoint[] getObservations()

Get the observed points.

Returns:

observed points

See Also:

addObservedPoint(double, double), addObservedPoint(double, double, double), addObservedPoint(WeightedObservedPoint)

clearObservations

public void clearObservations()

Remove all observations.

fit

public double[] fit(ParametricRealFunction f,
                    double[] initialGuess)
             throws FunctionEvaluationException,
                    OptimizationException,
                    java.lang.IllegalArgumentException

Fit a curve.

This method compute the coefficients of the curve that best fit the sample of observed points previously given through calls to the addObservedPoint method.

Parameters:

f - parametric function to fit

initialGuess - first guess of the function parameters

Returns:

fitted parameters

Throws:

FunctionEvaluationException - if the objective function throws one during the search

OptimizationException - if the algorithm failed to converge

java.lang.IllegalArgumentException - if the start point dimension is wrong