GaussianDistribution.h

Go to the documentation of this file.

// -*- C++ -*-4 1999/10/29 20:41:34 dugas

// PLearn (A C++ Machine Learning Library)
// Copyright (C) 2002 Pascal Vincent
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are met:
// 
//  1. Redistributions of source code must retain the above copyright
//     notice, this list of conditions and the following disclaimer.
// 
//  2. 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.
// 
//  3. The name of the authors may not be used to endorse or promote
//     products derived from this software without specific prior written
//     permission.
// 
// THIS SOFTWARE IS PROVIDED BY THE AUTHORS ``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 THE AUTHORS 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.
// 
// This file is part of the PLearn library. For more information on the PLearn
// library, go to the PLearn Web site at www.plearn.org


 

/* *******************************************************      
   * $Id: GaussianDistribution.h,v 1.9 2004/08/12 16:15:25 tihocan Exp $
   * This file is part of the PLearn library.
   ******************************************************* */


#ifndef GaussianDistribution_INC
#define GaussianDistribution_INC

#include "PDistribution.h"

namespace PLearn {
using namespace std;

  // This is a density estimation learner.
  // It uses a compact representation of a Gaussian, by keeping only the k 
  // top eigenvalues and associated eigenvectors of the covariance matrix.
  // All other eigenvalues are kept at the level of the k+1 th eigenvalue
  // Optionally, a constant sigma is first added to the diagonal of the covariance matrix.

  class GaussianDistribution: public PDistribution
  {
  public:    

    typedef PDistribution inherited;

    // Possibly "Learned" parameters
    Vec mu;
    Vec eigenvalues; 
    Mat eigenvectors; 

    // Build options
    int k; // maximum number of eigenvectors to keep
    real gamma; // The actual covariance matrix used will be VDV' + gamma.I where V'=eigenvectors and D=diag(eigenvalues)
    float ignore_weights_below; 
    
  public:
    GaussianDistribution();

    PLEARN_DECLARE_OBJECT(GaussianDistribution);
    void makeDeepCopyFromShallowCopy(CopiesMap& copies);

    virtual void forget();
    virtual void train();
    virtual real log_density(const Vec& x) const;

    virtual void resetGenerator(long g_seed) const;

    virtual void generate(Vec& x) const;

    virtual int inputsize() const;

  protected:
    static void declareOptions(OptionList& ol);

  };

  DECLARE_OBJECT_PTR(GaussianDistribution);

} // end of namespace PLearn

#endif

---