Kernel.h

Go to the documentation of this file.

// -*- C++ -*-

// PLearn (A C++ Machine Learning Library)
// Copyright (C) 1998 Pascal Vincent
// Copyright (C) 1999-2002 Pascal Vincent, Yoshua Bengio, Rejean Ducharme and University of Montreal
// Copyright (C) 2001-2002 Nicolas Chapados, Ichiro Takeuchi, Jean-Sebastien Senecal
// Copyright (C) 2002 Xiangdong Wang, Christian Dorion

// 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: Kernel.h,v 1.33 2004/07/23 13:42:52 tihocan Exp $
   * This file is part of the PLearn library.
   ******************************************************* */

#ifndef Kernel_INC
#define Kernel_INC

#include <plearn/base/Object.h>
#include <plearn/vmat/VMat.h>
//#include "PLMPI.h"

namespace PLearn {
using namespace std;

class Kernel: public Object
{

private:

  typedef Object inherited;

  mutable Vec evaluate_xi, evaluate_xj, k_xi_x;

  mutable bool lock_xi, lock_xj, lock_k_xi_x;
                
protected:

  VMat data; 
  int data_inputsize;
  int n_examples;

  static void declareOptions(OptionList& ol);

public:

  bool is_symmetric;
  int report_progress;
  VMat specify_dataset;

  Kernel(bool is__symmetric = true);

  PLEARN_DECLARE_ABSTRACT_OBJECT(Kernel);

  virtual real evaluate(const Vec& x1, const Vec& x2) const = 0; 


  virtual void setDataForKernelMatrix(VMat the_data);

  virtual void addDataForKernelMatrix(const Vec& newRow);


  virtual int dataInputsize() {
    return data_inputsize;
  }

  virtual int nExamples()
    { return n_examples; }

  virtual real evaluate_i_j(int i, int j) const; 

  virtual real evaluate_i_x(int i, const Vec& x, real squared_norm_of_x=-1) const; 

  virtual real evaluate_x_i(const Vec& x, int i, real squared_norm_of_x=-1) const; 

  virtual real evaluate_i_x_again(int i, const Vec& x, real squared_norm_of_x=-1, bool first_time = false) const;
  virtual real evaluate_x_i_again(const Vec& x, int i, real squared_norm_of_x=-1, bool first_time = false) const;

  virtual void computeGramMatrix(Mat K) const;

  virtual void setParameters(Vec paramvec); 
  virtual Vec getParameters() const; 


  void apply(VMat m1, VMat m2, Mat& result) const; 
  Mat apply(VMat m1, VMat m2) const; 
  void apply(VMat m, const Vec& x, Vec& result) const; 
  void apply(Vec x, VMat m, Vec& result) const; 

  void evaluate_all_i_x(const Vec& x, Vec& k_xi_x, real squared_norm_of_x=-1, int istart = 0) const;

  void evaluate_all_x_i(const Vec& x, Vec& k_x_xi, real squared_norm_of_x=-1, int istart = 0) const;

  inline real operator()(const Vec& x1, const Vec& x2) const
    { return evaluate(x1,x2); }

  bool hasData();

  inline VMat getData() {return this->data;}

  bool isInData(const Vec& x, int* i = 0) const;

  void computeNearestNeighbors(const Vec& x, Mat& k_xi_x_sorted, int knn) const;

  static TMat<int> computeKNNeighbourMatrixFromDistanceMatrix(const Mat& D, int knn, bool insure_self_first_neighbour=true, bool report_progress = false);

  static Mat computeNeighbourMatrixFromDistanceMatrix(const Mat& D, bool insure_self_first_neighbour=true, bool report_progress = false);

  Mat estimateHistograms(VMat d, real sameness_threshold, real minval, real maxval, int nbins) const;
  Mat estimateHistograms(Mat input_and_class, real minval, real maxval, int nbins) const;
  real test(VMat d, real threshold, real sameness_below_threshold, real sameness_above_threshold) const;
  virtual void build();
    //virtual void oldwrite(ostream& out) const;
    //virtual void oldread(istream& in);
  virtual ~Kernel();

  virtual void makeDeepCopyFromShallowCopy(CopiesMap& copies);

private:

  void build_();

};
DECLARE_OBJECT_PTR(Kernel);

class Ker: public PP<Kernel>
{
public:
  Ker() {}
  Ker(Kernel* v) :PP<Kernel>(v) {}
  Ker(const Ker& other) :PP<Kernel>(other) {}

  real operator()(const Vec& x1, const Vec& x2) const
    { return ptr->evaluate(x1,x2); }
};

DECLARE_OBJECT_PP(Ker, Kernel);

template <>
inline
void deepCopyField(Ker& field, CopiesMap& copies)
{
  if (field)
    field = static_cast<Kernel*>(field->deepCopy(copies));
}

// last column of data is supposed to be a classnum
// returns a matrix of (index1, index2, distance)
Mat findClosestPairsOfDifferentClass(int k, VMat data, Ker dist);


inline Array<Ker> operator&(const Ker& k1, const Ker& k2)
{ return Array<Ker>(k1,k2); }

/*!************
 * CostFunc *
 ************
*/

typedef Ker CostFunc;

/*!**********************************************************************
  FINANCIAL STUFF
  **********************************************************************
*/

typedef CostFunc ProfitFunc;


} // end of namespace PLearn

#endif

---