LearnerProcessedVMatrix.cc

Go to the documentation of this file.

// -*- C++ -*-

// LearnerProcessedVMatrix.cc
//
// Copyright (C) 2003  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: LearnerProcessedVMatrix.cc,v 1.6 2004/07/09 19:42:23 tihocan Exp $ 
   ******************************************************* */

#include "LearnerProcessedVMatrix.h"

namespace PLearn {
using namespace std;


PLEARN_IMPLEMENT_OBJECT(LearnerProcessedVMatrix, "ONE LINE DESCR", 
                        "LearnerProcessedVMatrix implements a VMatrix processed on the fly by a learner (which will optionally be first trained on the source vmatrix)");


LearnerProcessedVMatrix::LearnerProcessedVMatrix()
  :train_learner('-')
{
  build_();
}

void LearnerProcessedVMatrix::getNewRow(int i, const Vec& v) const
{
  static Vec sv;
  sv.resize(source.width());
  source->getRow(i,sv);
  int nin = source->inputsize();
  int nout = learner->outputsize();
  Vec input = sv.subVec(0,nin);
  Vec output = v.subVec(0,nout);
  learner->computeOutput(input,output);
  int rest = source.width()-nin;
  v.subVec(nout,rest) << sv.subVec(nin,rest);
}

void LearnerProcessedVMatrix::declareOptions(OptionList& ol)
{
  declareOption(ol, "source", &LearnerProcessedVMatrix::source, OptionBase::buildoption,
                "The source vmatrix whose inputs wil be porcessed by the learner.\n"
                "If present, the target and weight columns will be appended to the processed input in the resulting matrix.");
  
  declareOption(ol, "learner", &LearnerProcessedVMatrix::learner, OptionBase::buildoption,
                "The learner used to process the VMat's input.");

  declareOption(ol, "train_learner", &LearnerProcessedVMatrix::train_learner, OptionBase::buildoption, 
                "Indicates if the learner should be trained on the source, upon building, and if so on what part.\n"
                " '-': don't train \n"
                " 'S': supervised training using input and target (possibly weighted if weight is  present) \n"
                " 'U': unsupervised training using only input part (possibly weighted if weight is present). \n");

  // Now call the parent class' declareOptions
  inherited::declareOptions(ol);
}

void LearnerProcessedVMatrix::build_()
{
  if (source && learner) {
    switch(train_learner) {
    case '-':
       break;
    case 'S':
        learner->setTrainingSet(source);
        learner->setTrainStatsCollector(new VecStatsCollector());
        learner->train();
      break;
    case 'U':
      {
        VMat inputs = source.subMatColumns(0,source->inputsize());
        learner->setTrainingSet(inputs);
        learner->setTrainStatsCollector(new VecStatsCollector());
        learner->train();
      }
    }
    length_ = source->length();
    width_ = learner->outputsize() + source->targetsize() + source->weightsize();
  }
}

// ### Nothing to add here, simply calls build_
void LearnerProcessedVMatrix::build()
{
  inherited::build();
  build_();
}

void LearnerProcessedVMatrix::makeDeepCopyFromShallowCopy(map<const void*, void*>& copies)
{
  inherited::makeDeepCopyFromShallowCopy(copies);
  deepCopyField(source, copies);
  deepCopyField(learner, copies);
}

} // end of namespace PLearn

---