GaussMix.h

Go to the documentation of this file.

// -*- C++ -*-

// GaussMix.h
// 
// Copyright (C) 2004 Université de Montréal
// 
// 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: GaussMix.h,v 1.27 2004/07/21 20:22:36 tihocan Exp $ 
   ******************************************************* */

#ifndef GaussMix_INC
#define GaussMix_INC

#include "PDistribution.h"

namespace PLearn {
using namespace std;

class GaussMix: public PDistribution
{

private:

  typedef PDistribution inherited;

  Vec sample_row, log_likelihood_post;
  mutable Vec x_minus_mu_x, mu_target, log_likelihood_dens;

protected:

  // *********************
  // * protected options *
  // *********************

  Mat eigenvalues;
  TVec<Mat> eigenvectors;
  int D;
  Mat diags;
  Vec log_coeff;
  Vec log_p_j_x;
  Vec log_p_x_j_alphaj;
  Mat mu_y_x;
  int n_eigen_computed;
  int nsamples;
  Vec p_j_x;

  // Fields below are not options.

  TVec<Mat> cov_x;            
  TVec<Mat> cov_y_x;          
  Mat eigenvalues_x;          
  Mat eigenvalues_y_x;        
  TVec<Mat> eigenvectors_x;   
  TVec<Mat> eigenvectors_y_x; 
  TVec<Mat> full_cov;         
  TVec<Mat> y_x_mat;          

  Mat posteriors;

  Vec initial_weights;

  Mat updated_weights;

public:

  // ************************
  // * public build options *
  // ************************

  real alpha_min;
  real epsilon;
  int kmeans_iterations;
  int L;
  int n_eigen;
  real sigma_min;
  string type;

  // Currently learnt options, but may be build options in the future.
  Vec alpha;
  Mat mu;  
  Vec sigma;

  // ****************
  // * Constructors *
  // ****************

  GaussMix();

  // ******************
  // * Object methods *
  // ******************

protected:

  virtual void computeMeansAndCovariances();

  virtual real computeLogLikelihood(const Vec& y, int j, bool is_input = false) const;

  virtual void computePosteriors();

  virtual bool computeWeights();

  virtual void generateFromGaussian(Vec& s, int given_gaussian) const;

  virtual void precomputeStuff();

  virtual void replaceGaussian(int j);

  void resizeStuffBeforeTraining();

  void updateSampleWeights();

public:

  virtual void generate(Vec& s) const;

  virtual void resetGenerator(long g_seed) const;

private: 

  void build_();

protected: 

  static void declareOptions(OptionList& ol);

  void kmeans(VMat samples, int nclust, TVec<int> & clust_idx, Mat & clust, int maxit=9999);

public:

  virtual void forget();

  // simply calls inherited::build() then build_() 
  virtual void build();

  virtual void makeDeepCopyFromShallowCopy(map<const void*, void*>& copies);

  PLEARN_DECLARE_OBJECT(GaussMix);

  // ********************
  // * PLearner methods *
  // ********************

  virtual void train();

  // *************************
  // * PDistribution methods *
  // *************************

  virtual void setInput(const Vec& input) const;

  virtual void updateFromConditionalSorting();

  virtual real log_density(const Vec& y) const;

  virtual real survival_fn(const Vec& y) const;

  virtual real cdf(const Vec& y) const;

  virtual void expectation(Vec& mu) const;

  virtual void variance(Mat& cov) const;

  int getNEigenComputed() const;
  Mat getEigenvectors(int j) const;
  Vec getEigenvals(int j) const;

};

// Declares a few other classes and functions related to this class
DECLARE_OBJECT_PTR(GaussMix);

} // end of namespace PLearn

#endif

---