greedyThickThinning_tpl.h

Go to the documentation of this file.

/****************************************************************************
 *   This file is part of the aGrUM/pyAgrum library.                        *
 *                                                                          *
 *   Copyright (c) 2005-2026 by                                             *
 *       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 *
 *       - Christophe GONZALES(_at_AMU)                                     *
 *                                                                          *
 *   The aGrUM/pyAgrum library is free software; you can redistribute it    *
 *   and/or modify it under the terms of either :                           *
 *                                                                          *
 *    - the GNU Lesser General Public License as published by               *
 *      the Free Software Foundation, either version 3 of the License,      *
 *      or (at your option) any later version,                              *
 *    - the MIT license (MIT),                                              *
 *    - or both in dual license, as here.                                   *
 *                                                                          *
 *   (see https://agrum.gitlab.io/articles/dual-licenses-lgplv3mit.html)    *
 *                                                                          *
 *   This aGrUM/pyAgrum library is distributed in the hope that it will be  *
 *   useful, but WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,          *
 *   INCLUDING BUT NOT LIMITED TO THE WARRANTIES MERCHANTABILITY or FITNESS *
 *   FOR A PARTICULAR PURPOSE  AND NONINFRINGEMENT. IN NO EVENT SHALL THE   *
 *   AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER *
 *   LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,        *
 *   ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR  *
 *   OTHER DEALINGS IN THE SOFTWARE.                                        *
 *                                                                          *
 *   See LICENCES for more details.                                         *
 *                                                                          *
 *   SPDX-FileCopyrightText: Copyright 2005-2026                            *
 *       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 *
 *       - Christophe GONZALES(_at_AMU)                                     *
 *   SPDX-License-Identifier: LGPL-3.0-or-later OR MIT                      *
 *                                                                          *
 *   Contact  : info_at_agrum_dot_org                                       *
 *   homepage : http://agrum.gitlab.io                                      *
 *   gitlab   : https://gitlab.com/agrumery/agrum                           *
 *                                                                          *
 ****************************************************************************/
 
#pragma once
 

 
#include <agrum/BN/learning/greedyThickThinning.h>   // to ease IDE parser
#include <agrum/BN/learning/paramUtils/DAG2BNLearner.h>
#include <agrum/BN/learning/structureUtils/graphChange.h>
 
namespace gum {
 
  namespace learning {

    template < typename GRAPH_CHANGES_SELECTOR >
    DAG GreedyThickThinning::learnStructure(GRAPH_CHANGES_SELECTOR& selector, DAG dag) {
      initApproximationScheme();
 
      // Phase 1: thickening — arc additions only
      selector.useArcAdditions(true);
      selector.useArcDeletions(false);
      selector.useArcReversals(false);
      selector.useArcTriangleDeletions(false);
      selector.setGraph(dag);
 
      unsigned int nb_changes_applied = 1;
      double       delta_score;
 
      do {
        nb_changes_applied = 0;
        delta_score        = 0;
 
        try {
          const auto& change = selector.bestChange();
          delta_score        = selector.deltaScore(change, true);
 
          if ((delta_score > 0) && continueApproximationScheme(delta_score)) {
            selector.applyChange(change);
            nb_changes_applied = 1;
            updateApproximationScheme(nb_changes_applied);
          }
        } catch (NotFound&) {}
      } while (nb_changes_applied);
 
      // Phase 2: thinning — arc deletions (+ optional reversals), no additions
      selector.useArcAdditions(false);
      selector.useArcDeletions(true);
      selector.useArcReversals(_allowReversalsInThinPhase_);
      selector.useArcTriangleDeletions(false);
      selector.setGraph(dag);
 
      do {
        nb_changes_applied = 0;
        delta_score        = 0;
 
        try {
          const auto& change = selector.bestChange();
          delta_score        = selector.deltaScore(change, true);
 
          if ((delta_score > 0) && continueApproximationScheme(delta_score)) {
            selector.applyChange(change);
            nb_changes_applied = 1;
            updateApproximationScheme(nb_changes_applied);
          }
        } catch (NotFound&) {}
      } while (nb_changes_applied);
 
      stopApproximationScheme();
 
      selector.finalizeGraph(dag);
 
      return dag;
    }

    template < GUM_Numeric GUM_SCALAR, typename GRAPH_CHANGES_SELECTOR, typename PARAM_ESTIMATOR >
    BayesNet< GUM_SCALAR > GreedyThickThinning::learnBN(GRAPH_CHANGES_SELECTOR& selector,
                                                        PARAM_ESTIMATOR&        estimator,
                                                        DAG                     initial_dag) {
      return DAG2BNLearner::createBN< GUM_SCALAR >(estimator,
                                                   learnStructure(selector, initial_dag));
    }
 
  } /* namespace learning */
 
} /* namespace gum */