LevMarqFitter1D.h

Go to the documentation of this file.

// --------------------------------------------------------------------------
//                   OpenMS -- Open-Source Mass Spectrometry
// --------------------------------------------------------------------------
// Copyright The OpenMS Team -- Eberhard Karls University Tuebingen,
// ETH Zurich, and Freie Universitaet Berlin 2002-2012.
//
// This software is released under a three-clause BSD license:
//  * 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.
//  * Neither the name of any author or any participating institution
//    may be used to endorse or promote products derived from this software
//    without specific prior written permission.
// For a full list of authors, refer to the file AUTHORS.
// --------------------------------------------------------------------------
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
// ARE DISCLAIMED. IN NO EVENT SHALL ANY OF THE AUTHORS OR THE CONTRIBUTING
// INSTITUTIONS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
// OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
// WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
// OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
// ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
// --------------------------------------------------------------------------
// $Maintainer: Clemens Groepl $
// $Authors: $
// --------------------------------------------------------------------------

#ifndef OPENMS_TRANSFORMATIONS_FEATUREFINDER_LEVMARQFITTER1D_H
#define OPENMS_TRANSFORMATIONS_FEATUREFINDER_LEVMARQFITTER1D_H

#include <OpenMS/TRANSFORMATIONS/FEATUREFINDER/Fitter1D.h>

#include <gsl/gsl_rng.h> // gsl random number generators
#include <gsl/gsl_randist.h> // gsl random number distributions
#include <gsl/gsl_vector.h> // gsl vector and matrix definitions
#include <gsl/gsl_multifit_nlin.h> // gsl multidimensional fitting
#include <gsl/gsl_blas.h> // gsl linear algebra stuff

namespace OpenMS
{

  class OPENMS_DLLAPI LevMarqFitter1D :
    public Fitter1D
  {

public:

    typedef std::vector<double> ContainerType;

    LevMarqFitter1D() :
      Fitter1D()
    {
      this->defaults_.setValue("max_iteration", 500, "Maximum number of iterations using by Levenberg-Marquardt algorithm.", StringList::create("advanced"));
      this->defaults_.setValue("deltaAbsError", 0.0001, "Absolute error used by the Levenberg-Marquardt algorithm.", StringList::create("advanced"));
      this->defaults_.setValue("deltaRelError", 0.0001, "Relative error used by the Levenberg-Marquardt algorithm.", StringList::create("advanced"));
    }

    LevMarqFitter1D(const LevMarqFitter1D & source) :
      Fitter1D(source),
      max_iteration_(source.max_iteration_),
      abs_error_(source.abs_error_),
      rel_error_(source.rel_error_)
    {
    }

    virtual ~LevMarqFitter1D()
    {
    }

    virtual LevMarqFitter1D & operator=(const LevMarqFitter1D & source)
    {
      if (&source == this) return *this;

      Fitter1D::operator=(source);
      max_iteration_ = source.max_iteration_;
      abs_error_ = source.abs_error_;
      rel_error_ = source.rel_error_;

      return *this;
    }

protected:

    Int gsl_status_;
    bool symmetric_;
    Int max_iteration_;

    CoordinateType abs_error_;
    CoordinateType rel_error_;

    virtual void printState_(Int iter, gsl_multifit_fdfsolver * s) = 0;

    const String getGslStatus_()
    {
      return gsl_strerror(gsl_status_);
    }

    void optimize_(const RawDataArrayType & set, Int num_params, CoordinateType x_init[],
                   Int (* residual)(const gsl_vector * x, void * params, gsl_vector * f),
                   Int (* jacobian)(const gsl_vector * x, void * params, gsl_matrix * J),
                   Int (* evaluate)(const gsl_vector * x, void * params, gsl_vector * f, gsl_matrix * J),
                   void * advanced_params
                   )
    {

      const gsl_multifit_fdfsolver_type * T;
      gsl_multifit_fdfsolver * s;

      Int status;
      Int iter = 0;
      const UInt n = (UInt)set.size();

      // number of parameters to be optimized
      UInt p = num_params;

      // gsl always expects N>=p or default gsl error handler invoked,
      // cause Jacobian be rectangular M x N with M>=N
      if (n < p) throw Exception::UnableToFit(__FILE__, __LINE__, __PRETTY_FUNCTION__, "UnableToFit-FinalSet", "Skipping feature, gsl always expects N>=p");

      // allocate space for a covariance matrix of size p by p
      gsl_matrix * covar = gsl_matrix_alloc(p, p);
      gsl_multifit_function_fdf f;

      gsl_vector_view x = gsl_vector_view_array(x_init, p);

      gsl_rng_env_setup();

      // set up the function to be fit
      f.f = (residual);     // the function of residuals
      f.df = (jacobian);     // the gradient of this function
      f.fdf = (evaluate);     // combined function and gradient
      f.n = set.size();     // number of points in the data set
      f.p = p;     // number of parameters in the fit function
      f.params = advanced_params;     // // structure with the data and error bars

      T = gsl_multifit_fdfsolver_lmsder;
      s = gsl_multifit_fdfsolver_alloc(T, n, p);
      gsl_multifit_fdfsolver_set(s, &f, &x.vector);

#ifdef DEBUG_FEATUREFINDER
      printState_(iter, s);
#endif

      // this is the loop for fitting
      do
      {
        iter++;

        // perform a single iteration of the fitting routine
        status = gsl_multifit_fdfsolver_iterate(s);

#ifdef DEBUG_FEATUREFINDER
        // customized routine to print out current parameters
        printState_(iter, s);
#endif

        /* check if solver is stuck */
        if (status) break;

        // test for convergence with an absolute and relative error
        status = gsl_multifit_test_delta(s->dx, s->x, abs_error_, rel_error_);
      }
      while (status == GSL_CONTINUE && iter < max_iteration_);

      // This function uses Jacobian matrix J to compute the covariance matrix of the best-fit parameters, covar.
      // The parameter epsrel (0.0) is used to remove linear-dependent columns when J is rank deficient.
      gsl_multifit_covar(s->J, 0.0, covar);

#ifdef DEBUG_FEATUREFINDER
      gsl_matrix_fprintf(stdout, covar, "covar %g");
#endif

#define FIT(i) gsl_vector_get(s->x, i)
#define ERR(i) sqrt(gsl_matrix_get(covar, i, i))

      // Set GSl status
      gsl_status_ = status;

#ifdef DEBUG_FEATUREFINDER
      {
        // chi-squared value
        DoubleReal chi = gsl_blas_dnrm2(s->f);
        DoubleReal dof = n - p;
        DoubleReal c = GSL_MAX_DBL(1, chi / sqrt(dof));

        printf("chisq/dof = %g\n", pow(chi, 2.0) / dof);

        for (Size i = 0; i < p; ++i)
        {
          std::cout << i;
          printf(".Parameter = %.5f +/- %.5f\n", FIT(i), c * ERR(i));
        }
      }
#endif

      // set optimized parameters
      for (Size i = 0; i < p; ++i)
      {
        x_init[i] = FIT(i);
      }

      gsl_multifit_fdfsolver_free(s);
      gsl_matrix_free(covar);

    }

    void updateMembers_()
    {
      Fitter1D::updateMembers_();
      max_iteration_ = this->param_.getValue("max_iteration");
      abs_error_ = this->param_.getValue("deltaAbsError");
      rel_error_ = this->param_.getValue("deltaRelError");
    }

  };
}

#endif // OPENMS_TRANSFORMATIONS_FEATUREFINDER_LEVMARQFITTER1D_H