CHAIDTree.h

Go to the documentation of this file.

/*
 * 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.
 */


#ifndef _CHAIDTree_H__
#define _CHAIDTree_H__

#include <shogun/lib/config.h>

#include <shogun/multiclass/tree/TreeMachine.h>
#include <shogun/multiclass/tree/CHAIDTreeNodeData.h>
#include <shogun/features/DenseFeatures.h>

namespace shogun
{

class CCHAIDTree : public CTreeMachine<CHAIDTreeNodeData>
{
public:
    CCHAIDTree();

    CCHAIDTree(int32_t dependent_vartype);

    CCHAIDTree(int32_t dependent_vartype, SGVector<int32_t> feature_types, int32_t num_breakpoints=0);

    virtual ~CCHAIDTree();

    virtual const char* get_name() const { return "CHAIDTree"; }

    virtual EProblemType get_machine_problem_type() const;

    virtual bool is_label_valid(CLabels* lab) const;

    virtual CMulticlassLabels* apply_multiclass(CFeatures* data=NULL);

    virtual CRegressionLabels* apply_regression(CFeatures* data=NULL);

    void set_weights(SGVector<float64_t> w);

    SGVector<float64_t> get_weights() const;

    void clear_weights();

    void set_feature_types(SGVector<int32_t> ft);

    SGVector<int32_t> get_feature_types() const;

    void clear_feature_types();

    void set_dependent_vartype(int32_t var);

    int32_t get_dependent_vartype() const { return m_dependent_vartype; }

    void set_max_tree_depth(int32_t d) { m_max_tree_depth=d; }

    int32_t get_specified_max_tree_depth() const { return m_max_tree_depth; }

    void set_min_node_size(int32_t size) { m_min_node_size=size; }

    int32_t get_min_node_size() const { return m_min_node_size; }

    void set_alpha_merge(float64_t a) { m_alpha_merge=a; }

    float64_t get_alpha_merge() const { return m_alpha_merge; }

    void set_alpha_split(float64_t a) { m_alpha_split=a; }

    float64_t get_alpha_split() const { return m_alpha_split; }

    void set_num_breakpoints(int32_t b) { m_num_breakpoints=b; }

    float64_t get_num_breakpoints() const { return m_num_breakpoints; }

protected:
    virtual bool train_machine(CFeatures* data=NULL);

private:
    CTreeMachineNode<CHAIDTreeNodeData>* CHAIDtrain(CFeatures* data, SGVector<float64_t> weights, CLabels* labels, int32_t level);

    SGVector<int32_t> merge_categories_ordinal(SGVector<float64_t> feats, SGVector<float64_t> labels,
                                SGVector<float64_t> weights, float64_t &pv);

    SGVector<int32_t> merge_categories_nominal(SGVector<float64_t> feats, SGVector<float64_t> labels,
                                SGVector<float64_t> weights, float64_t &pv);

    CLabels* apply_tree(CFeatures* data);

    CLabels* apply_from_current_node(SGMatrix<float64_t> fmat, node_t* current);

    bool handle_missing_ordinal(SGVector<int32_t> cat, SGVector<float64_t> feats, SGVector<float64_t> labels, SGVector<float64_t> weights);

    float64_t adjusted_p_value(float64_t p_value, int32_t inum_cat, int32_t fnum_cat, int32_t ft, bool is_missing);

    float64_t p_value(SGVector<float64_t> feat, SGVector<float64_t> labels, SGVector<float64_t> weights);

    float64_t anova_f_statistic(SGVector<float64_t> feat, SGVector<float64_t> labels, SGVector<float64_t> weights, int32_t &r);

    float64_t likelihood_ratio_statistic(SGVector<float64_t> feat, SGVector<float64_t> labels, SGVector<float64_t> weights,
                                                     int32_t &r, int32_t &c);

    float64_t pchi2_statistic(SGVector<float64_t> feat, SGVector<float64_t> labels, SGVector<float64_t> weights, int32_t &r, int32_t &c);

    SGMatrix<float64_t> expected_cf_row_effects_model(SGMatrix<int32_t> ct, SGMatrix<float64_t> wt, SGVector<float64_t> score);

    SGMatrix<float64_t> expected_cf_indep_model(SGMatrix<int32_t> ct, SGMatrix<float64_t> wt);

    float64_t sum_of_squared_deviation(SGVector<float64_t> lab, SGVector<float64_t> weights, float64_t &mean);

    bool continuous_to_ordinal(CDenseFeatures<float64_t>* feats);

    void modify_data_matrix(CDenseFeatures<float64_t>* feats);

    void init();

public:
    static const float64_t MISSING;

private:
    SGVector<int32_t> m_feature_types;

    SGVector<float64_t> m_weights;

    bool m_weights_set;

    int32_t m_dependent_vartype;

    int32_t m_max_tree_depth;

    int32_t m_min_node_size;

    float64_t m_alpha_merge;

    float64_t m_alpha_split;

    SGMatrix<float64_t> m_cont_breakpoints;

    int32_t m_num_breakpoints;

};
} /* namespace shogun */

#endif /* _CHAIDTree_H__ */