ToBagSplitter.cc

Go to the documentation of this file.

// -*- C++ -*-

// ToBagSplitter.cc
//
// Copyright (C) 2004 Olivier Delalleau 
// 
// 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: ToBagSplitter.cc,v 1.3 2004/07/21 16:30:55 chrish42 Exp $ 
   ******************************************************* */

// Authors: Olivier Delalleau

#include "SelectRowsVMatrix.h"
#include <plearn/var/SumOverBagsVariable.h>  
#include "ToBagSplitter.h"

namespace PLearn {
using namespace std;

// ToBagSplitter //
ToBagSplitter::ToBagSplitter() 
  : Splitter(),
    expected_size_of_bag(10)
{}

PLEARN_IMPLEMENT_OBJECT(ToBagSplitter,
    "A Splitter that makes any existing splitter operate on bags only.",
    "The dataset provided must contain bag information, as described in\n"
    "SumOverBagsVariable");

// declareOptions //
void ToBagSplitter::declareOptions(OptionList& ol)
{
   declareOption(ol, "expected_size_of_bag", &ToBagSplitter::expected_size_of_bag, OptionBase::buildoption,
       "The expected size of each bag. It is not compulsory to change this option.");

   declareOption(ol, "sub_splitter", &ToBagSplitter::sub_splitter, OptionBase::buildoption,
       "The underlying splitter we want to make operate on bags.");

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

// build //
void ToBagSplitter::build()
{
  inherited::build();
  build_();
}

// build_ //
void ToBagSplitter::build_()
{
  if (dataset) {
    // Prepare the bags index list.
    int max_ninstances = 1;
    // The first column in bags_store gives the number of instances in the bag,
    // and the following columns give the indices of the corresponding rows in
    // the original dataset.
    Mat bags_store(dataset->length() / expected_size_of_bag + 1, expected_size_of_bag + 1);
    int num_bag = 0;
    int num_instance = 0;
    int bag_signal_column = dataset->inputsize() + dataset->targetsize() - 1; // Bag signal in the last target column.
    for (int i = 0; i < dataset->length(); i++) {
      if (num_instance + 1 >= bags_store.width()) {
        if (num_instance > 10*(expected_size_of_bag+1))
          PLERROR("ToBagSplitter: found bag size (%d) more than 10 times bigger than expected_size_of_bag (%d)!\n",
                  num_instance,expected_size_of_bag);
        // Need to resize bags_store.
        bags_store.resize(bags_store.length(), bags_store.width() * 2);
      }
      if (num_instance >= max_ninstances) {
        max_ninstances = num_instance + 1;
      }
      bags_store(num_bag, num_instance + 1) = i;
      num_instance++;
      if (int(dataset->get(i, bag_signal_column)) & SumOverBagsVariable::TARGET_COLUMN_LAST) {
        // Last element of a bag.
        bags_store(num_bag, 0) = num_instance; // Store the number of instances in this bag.
        num_bag++;
        num_instance = 0;
        if (num_bag >= bags_store.length()) {
          // Need to resize bags_store.
          bags_store.resize(bags_store.length() * 2, bags_store.width());
        }
      }
    }
    // Resize to the minimum size needed.
    bags_store.resize(num_bag, max_ninstances + 1);
    bags_index = VMat(bags_store);
    // Provide this index to the sub_splitter.
    sub_splitter->setDataSet(bags_index);
  }
}

// getSplit //
TVec<VMat> ToBagSplitter::getSplit(int k)
{
  // ### Build and return the kth split 
  TVec<VMat> sub_splits = sub_splitter->getSplit(k);
  TVec<VMat> result;
  for (int i = 0; i < sub_splits.length(); i++) {
    // Get the list of corresponding indices in the original dataset.
    Mat indices = sub_splits[i].toMat();
    // Turn it into a TVec<int>.
    TVec<int> indices_int;
    for (int j = 0; j < indices.length(); j++) {
      for (int k = 0; k < indices(j, 0); k++) {
        int indice = int(indices(j, k + 1));
        indices_int.append(indice);
      }
    }
    result.append(new SelectRowsVMatrix(dataset, indices_int));
  }
  return result;
}

// makeDeepCopyFromShallowCopy //
void ToBagSplitter::makeDeepCopyFromShallowCopy(map<const void*, void*>& copies)
{
  Splitter::makeDeepCopyFromShallowCopy(copies);

  // ### Call deepCopyField on all "pointer-like" fields 
  // ### that you wish to be deepCopied rather than 
  // ### shallow-copied.
  // ### ex:
  // deepCopyField(trainvec, copies);

  // ### Remove this line when you have fully implemented this method.
  PLERROR("ToBagSplitter::makeDeepCopyFromShallowCopy not fully (correctly) implemented yet!");
}

// nSetsPerSplit //
int ToBagSplitter::nSetsPerSplit() const
{
  // ### Return the number of sets per split
  return sub_splitter->nSetsPerSplit();
}

// nsplits //
int ToBagSplitter::nsplits() const
{
  return sub_splitter->nsplits();
}

// setDataSet //
void ToBagSplitter::setDataSet(VMat the_dataset) {
  inherited::setDataSet(the_dataset);
  // Need to recompute the bags index.
  build();
}

} // end of namespace PLearn

---