org.apache.commons.math.estimation

Class LevenbergMarquardtEstimator

java.lang.Object

org.apache.commons.math.estimation.AbstractEstimator

org.apache.commons.math.estimation.LevenbergMarquardtEstimator

All Implemented Interfaces:

java.io.Serializable, Estimator

Deprecated.

as of 2.0, everything in package org.apache.commons.math.estimation has been deprecated and replaced by package org.apache.commons.math.optimization.general

@Deprecated
public class LevenbergMarquardtEstimator
extends AbstractEstimator
implements java.io.Serializable

This class solves a least squares problem.

This implementation should work even for over-determined systems (i.e. systems having more variables than equations). Over-determined systems are solved by ignoring the variables which have the smallest impact according to their jacobian column norm. Only the rank of the matrix and some loop bounds are changed to implement this.

The resolution engine is a simple translation of the MINPACK lmder routine with minor changes. The changes include the over-determined resolution and the Q.R. decomposition which has been rewritten following the algorithm described in the P. Lascaux and R. Theodor book Analyse numérique matricielle appliquée à l'art de l'ingénieur, Masson 1986.

The authors of the original fortran version are:

• Argonne National Laboratory. MINPACK project. March 1980
• Burton S. Garbow
• Kenneth E. Hillstrom
• Jorge J. More

The redistribution policy for MINPACK is available here, for convenience, it is reproduced below.

Minpack Copyright Notice (1999) University of Chicago. All rights reserved

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. The end-user documentation included with the redistribution, if any, must include the following acknowledgment: This product includes software developed by the University of Chicago, as Operator of Argonne National Laboratory. Alternately, this acknowledgment may appear in the software itself, if and wherever such third-party acknowledgments normally appear. WARRANTY DISCLAIMER. THE SOFTWARE IS SUPPLIED "AS IS" WITHOUT WARRANTY OF ANY KIND. THE COPYRIGHT HOLDER, THE UNITED STATES, THE UNITED STATES DEPARTMENT OF ENERGY, AND THEIR EMPLOYEES: (1) DISCLAIM ANY WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE OR NON-INFRINGEMENT, (2) DO NOT ASSUME ANY LEGAL LIABILITY OR RESPONSIBILITY FOR THE ACCURACY, COMPLETENESS, OR USEFULNESS OF THE SOFTWARE, (3) DO NOT REPRESENT THAT USE OF THE SOFTWARE WOULD NOT INFRINGE PRIVATELY OWNED RIGHTS, (4) DO NOT WARRANT THAT THE SOFTWARE WILL FUNCTION UNINTERRUPTED, THAT IT IS ERROR-FREE OR THAT ANY ERRORS WILL BE CORRECTED. LIMITATION OF LIABILITY. IN NO EVENT WILL THE COPYRIGHT HOLDER, THE UNITED STATES, THE UNITED STATES DEPARTMENT OF ENERGY, OR THEIR EMPLOYEES: BE LIABLE FOR ANY INDIRECT, INCIDENTAL, CONSEQUENTIAL, SPECIAL OR PUNITIVE DAMAGES OF ANY KIND OR NATURE, INCLUDING BUT NOT LIMITED TO LOSS OF PROFITS OR LOSS OF DATA, FOR ANY REASON WHATSOEVER, WHETHER SUCH LIABILITY IS ASSERTED ON THE BASIS OF CONTRACT, TORT (INCLUDING NEGLIGENCE OR STRICT LIABILITY), OR OTHERWISE, EVEN IF ANY OF SAID PARTIES HAS BEEN WARNED OF THE POSSIBILITY OF SUCH LOSS OR DAMAGES.

Since:

1.2

See Also:

Serialized Form

Field Summary

Fields inherited from class org.apache.commons.math.estimation.AbstractEstimator

DEFAULT_MAX_COST_EVALUATIONS

Constructor Summary

Constructors

Constructor and Description
LevenbergMarquardtEstimator() Deprecated. Build an estimator for least squares problems.

Method Summary

Modifier and Type	Method and Description
void	estimate(EstimationProblem problem) Deprecated. Solve an estimation problem using the Levenberg-Marquardt algorithm.
void	setCostRelativeTolerance(double costRelativeTolerance) Deprecated. Set the desired relative error in the sum of squares.
void	setInitialStepBoundFactor(double initialStepBoundFactor) Deprecated. Set the positive input variable used in determining the initial step bound.
void	setOrthoTolerance(double orthoTolerance) Deprecated. Set the desired max cosine on the orthogonality.
void	setParRelativeTolerance(double parRelativeTolerance) Deprecated. Set the desired relative error in the approximate solution parameters.

Methods inherited from class org.apache.commons.math.estimation.AbstractEstimator

getChiSquare, getCostEvaluations, getCovariances, getJacobianEvaluations, getRMS, guessParametersErrors, setMaxCostEval

Methods inherited from class java.lang.Object

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

Constructor Detail

LevenbergMarquardtEstimator

public LevenbergMarquardtEstimator()

Deprecated.

Build an estimator for least squares problems.

The default values for the algorithm settings are:

• initial step bound factor: 100.0
• maximal cost evaluations: 1000
• cost relative tolerance: 1.0e-10
• parameters relative tolerance: 1.0e-10
• orthogonality tolerance: 1.0e-10

Method Detail

setInitialStepBoundFactor

public void setInitialStepBoundFactor(double initialStepBoundFactor)

Deprecated.

Set the positive input variable used in determining the initial step bound. This bound is set to the product of initialStepBoundFactor and the euclidean norm of diag*x if nonzero, or else to initialStepBoundFactor itself. In most cases factor should lie in the interval (0.1, 100.0). 100.0 is a generally recommended value

Parameters:

initialStepBoundFactor - initial step bound factor

See Also:

estimate(org.apache.commons.math.estimation.EstimationProblem)

setCostRelativeTolerance

public void setCostRelativeTolerance(double costRelativeTolerance)

Deprecated.

Set the desired relative error in the sum of squares.

Parameters:

costRelativeTolerance - desired relative error in the sum of squares

See Also:

estimate(org.apache.commons.math.estimation.EstimationProblem)

setParRelativeTolerance

public void setParRelativeTolerance(double parRelativeTolerance)

Deprecated.

Set the desired relative error in the approximate solution parameters.

Parameters:

parRelativeTolerance - desired relative error in the approximate solution parameters

See Also:

estimate(org.apache.commons.math.estimation.EstimationProblem)

setOrthoTolerance

public void setOrthoTolerance(double orthoTolerance)

Deprecated.

Set the desired max cosine on the orthogonality.

Parameters:

orthoTolerance - desired max cosine on the orthogonality between the function vector and the columns of the jacobian

See Also:

estimate(org.apache.commons.math.estimation.EstimationProblem)

estimate

public void estimate(EstimationProblem problem)
              throws EstimationException

Deprecated.

Solve an estimation problem using the Levenberg-Marquardt algorithm.

The algorithm used is a modified Levenberg-Marquardt one, based on the MINPACK lmder routine. The algorithm settings must have been set up before this method is called with the setInitialStepBoundFactor(double), AbstractEstimator.setMaxCostEval(int), setCostRelativeTolerance(double), setParRelativeTolerance(double) and setOrthoTolerance(double) methods. If these methods have not been called, the default values set up by the constructor will be used.

The authors of the original fortran function are:

• Argonne National Laboratory. MINPACK project. March 1980
• Burton S. Garbow
• Kenneth E. Hillstrom
• Jorge J. More

Luc Maisonobe did the Java translation.

Specified by:

estimate in interface Estimator

Specified by:

estimate in class AbstractEstimator

Parameters:

problem - estimation problem to solve

Throws:

EstimationException - if convergence cannot be reached with the specified algorithm settings or if there are more variables than equations

See Also:

setInitialStepBoundFactor(double), setCostRelativeTolerance(double), setParRelativeTolerance(double), setOrthoTolerance(double)