KDTree.cpp

浏览该文件的文档.

/*
 * Copyright (c) The Shogun Machine Learning Toolbox
 * Written (w) 2014 Parijat Mazumdar
 * All rights reserved.
 *
 * 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.
 *
 * 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.
 *
 * The views and conclusions contained in the software and documentation are those
 * of the authors and should not be interpreted as representing official policies,
 * either expressed or implied, of the Shogun Development Team.
 */

#include <shogun/multiclass/tree/KDTree.h>

using namespace shogun;

CKDTree::CKDTree(int32_t leaf_size, EDistanceType d)
: CNbodyTree(leaf_size,d)
{
}

CKDTree::~CKDTree()
{
}

float64_t CKDTree::min_dist(bnode_t* node,float64_t* feat, int32_t dim)
{
    float64_t dist=0;
    for (int32_t i=0;i<dim;i++)
    {
        float64_t dim_dist=(node->data.bbox_lower[i]-feat[i])+CMath::abs(feat[i]-node->data.bbox_lower[i]);
        dim_dist+=(feat[i]-node->data.bbox_upper[i])+CMath::abs(feat[i]-node->data.bbox_upper[i]);
        dist+=add_dim_dist(0.5*dim_dist);
    }

    return actual_dists(dist);
}

float64_t CKDTree::min_dist_dual(bnode_t* nodeq, bnode_t* noder)
{
    SGVector<float64_t> nodeq_lower=nodeq->data.bbox_lower;
    SGVector<float64_t> nodeq_upper=nodeq->data.bbox_upper;
    SGVector<float64_t> noder_lower=noder->data.bbox_lower;
    SGVector<float64_t> noder_upper=noder->data.bbox_upper;
    float64_t dist=0;
    for(int32_t i=0;i<noder_lower.vlen;i++)
    {
        float64_t d1=nodeq_lower[i]-noder_upper[i];
        float64_t d2=noder_lower[i]-nodeq_upper[i];
        dist+=add_dim_dist(0.5*(d1+CMath::abs(d1)+d2+CMath::abs(d2)));
    }

    return actual_dists(dist);
}

float64_t CKDTree::max_dist_dual(bnode_t* nodeq, bnode_t* noder)
{
    SGVector<float64_t> nodeq_lower=nodeq->data.bbox_lower;
    SGVector<float64_t> nodeq_upper=nodeq->data.bbox_upper;
    SGVector<float64_t> noder_lower=noder->data.bbox_lower;
    SGVector<float64_t> noder_upper=noder->data.bbox_upper;
    float64_t dist=0;
    for(int32_t i=0;i<noder_lower.vlen;i++)
    {
        float64_t d1=CMath::abs(nodeq_lower[i]-noder_upper[i]);
        float64_t d2=CMath::abs(noder_lower[i]-nodeq_upper[i]);
        dist+=add_dim_dist(CMath::max(d1,d2));
    }

    return actual_dists(dist);
}

void CKDTree::min_max_dist(float64_t* pt, bnode_t* node, float64_t &lower,float64_t &upper, int32_t dim)
{
    lower=0;
    upper=0;
    for(int32_t i=0;i<dim;i++)
    {
        float64_t low_dist=node->data.bbox_lower[i]-pt[i];
        float64_t high_dist=pt[i]-node->data.bbox_upper[i];
        lower+=add_dim_dist(0.5*(low_dist+CMath::abs(low_dist)+high_dist+CMath::abs(high_dist)));
        upper+=add_dim_dist(CMath::max(CMath::abs(low_dist),CMath::abs(high_dist)));
    }

    lower=actual_dists(lower);
    upper=actual_dists(upper);
}

void CKDTree::init_node(bnode_t* node, index_t start, index_t end)
{
    SGVector<float64_t> upper_bounds(m_data.num_rows);
    SGVector<float64_t> lower_bounds(m_data.num_rows);

    for (int32_t i=0;i<m_data.num_rows;i++)
    {
        upper_bounds[i]=m_data(i,m_vec_id[start]);
        lower_bounds[i]=m_data(i,m_vec_id[start]);
        for (int32_t j=start+1;j<=end;j++)
        {
            upper_bounds[i]=CMath::max(upper_bounds[i],m_data(i,m_vec_id[j]));
            lower_bounds[i]=CMath::min(lower_bounds[i],m_data(i,m_vec_id[j]));
        }
    }

    float64_t radius=0;
    for (int32_t i=0;i<m_data.num_rows;i++)
        radius=CMath::max(radius,upper_bounds[i]-lower_bounds[i]);

    node->data.bbox_upper=upper_bounds;
    node->data.bbox_lower=lower_bounds;
    node->data.radius=0.5*radius;
    node->data.start_idx=start;
    node->data.end_idx=end;
}