binarydecisiontree.h

Go to the documentation of this file.

/*

Copyright (c) 2003, Cornell University
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:

   - Redistributions of source code must retain the above copyright notice,
       this list of conditions and the following disclaimer.
   - 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.
   - Neither the name of Cornell University nor the names of its
       contributors may be used to endorse or promote products derived from
       this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "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 COPYRIGHT OWNER OR CONTRIBUTORS 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.

*/

// -*- C++ -*-

#ifndef _CLUS_BINARYDECISIONTREE_H_
#define _CLUS_BINARYDECISIONTREE_H_

#include "machine.h"
#include "binarydecisiontreenode.h"
#include "dctraingen.h"
#include <iostream>

using namespace std;

namespace CLUS
{

/// Implements the binary decision tree
template< class T_Splitter >
class BinaryDecisionTree : public Machine
{
protected:
    BinaryDecisionTreeNode< T_Splitter >* root;

    /// vector of discrete domain sizes
    const Vector<int>& dDomainSize;

    /// number of discrete split variables
    int dsplitDim;
    
    /// number of continuous split variables
    int csplitDim;

    /// the minimum mass (sum of weights) to continue splitting
    int minMass; 

public:
    BinaryDecisionTree(const Vector<int>& DDomainSize,
                       int CsplitDim):
            Machine(CsplitDim,1),dDomainSize(DDomainSize),
            dsplitDim(DDomainSize.dim()-1),
            csplitDim(CsplitDim)
    {
        minMass=10;
        root=NULL;
    }

    ~BinaryDecisionTree(void)
    {
        if (root!=0)
            delete root;
    }

    virtual int InDim(void)
    {
        return dsplitDim+csplitDim;
    }
    
    virtual string TypeName(void)
    {
        return string("BinaryDecisionTree");
    }
    
    virtual void Infer(void)
    {
        if (root==0)
            return;
        // translate the first dsplitDim inputs into ints
        int Dvars[MAX_VARIABLES];
        for(int i=0; i<dsplitDim; i++)
        {
            Dvars[i]=(int)(*input)[i];
        }

        double Cvars[MAX_VARIABLES];
        for (int i=0; i<csplitDim; i++)
        {
            Cvars[i]=(*input)[i+dsplitDim];
        }

        output[0]=root->Infer(Dvars,Cvars);
    }
    
    virtual void Identify(void)
    {
        const Matrix<double>& ctrainData = training->GetTrainingData();
        const Matrix<int>& dtrainData = dynamic_cast< DCTrainingData* >( training )
                                        -> GetDiscreteTrainingData();

        int M=ctrainData.num_rows();

        // create the root and give it the Id 1
        root = new BinaryDecisionTreeNode< T_Splitter >
               (1, dDomainSize, csplitDim);

        do
        {
            root->StartLearningEpoch();
            for(int i=0; i<M; i++)
            {
                int classLbl=dtrainData[i][dsplitDim];
#ifdef DEBUG_PRINT

                cout << " i=" << i << " CL=" << classLbl << endl;
#endif

                root->LearnSample(dtrainData[i],ctrainData[i],classLbl);
            }
        }
        while (root->StopLearningEpoch(minMass));

        cout << "End Learning" << endl;
        // PrintSizeTree();
    }

    virtual void Prune(void)
    {
        const Matrix<double>& ctrainData = pruning->GetTrainingData();
        const Matrix<int>& dtrainData = dynamic_cast< DCTrainingData* >( pruning )
                                        -> GetDiscreteTrainingData();

        int M=ctrainData.num_rows();

        root->InitializePruningStatistics();
        for(int i=0; i<M; i++)
        {
            int classLbl=dtrainData[i][dsplitDim];
            root->UpdatePruningStatistics(dtrainData[i], ctrainData[i], classLbl);
        }
        root->FinalizePruningStatistics();

        // now cut the tree to the right size
        double cost=root->PruneSubtree();
        cout << "RMSN after pruning is:" << cost/M << endl;
        // PrintSizeTree();
    }
    
    virtual int SetOption(char* name, char* val)
    {
        if (strcmp(name,"MinMass")==0)
            minMass = atoi(val);
        else
            return Machine::SetOption(name,val);
        return 1;
    }

    virtual void SaveToStream(ostream& out)
    {
        out << TypeName() << " (  " << dsplitDim << " " << csplitDim << " ) { "  << endl;

        out << "[ ";
        for(int i=0; i<dsplitDim; i++)
            out << dDomainSize[i] << " ";
        out << "]" << endl;

        root->SaveToStream(out);

        out << '}' << endl;
    }
};

}

#endif // _CLUS_BINARYDECISIONTREE_H_

---