result_tracker_regularizer.hpp

Go to the documentation of this file.

#ifndef FSTRESULTTRACKERREGULARIZER_H
#define FSTRESULTTRACKERREGULARIZER_H
 /* 
=========================================================================
Copyright and license:
  * FST4 software (with the exception of any bundled or externally linked
    third-party components) is Copyright (C) 2000-present Institute of
    Information Theory and Automation of the Czech Academy of Sciences (UTIA).
  * FST4 is free software: you can redistribute it and/or modify it under
    the terms of the GNU Affero General Public License as published by the
    Free Software Foundation, version 3. The full text of the license is in
    the file LICENSE in the distribution root and at
    https://www.gnu.org/licenses/agpl-3.0.html.
  * FST4 is distributed in the hope that it will be useful, but WITHOUT ANY
    WARRANTY; without even the implied warranty of MERCHANTABILITY or
    FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public
    License for more details.
  * Closed-source use without the AGPL's copyleft obligations requires a
    commercial license from UTIA; in case of interest write to
    utia@utia.cas.cz. LICENSING.md in the distribution root explains both
    options, the contributor license agreement that every contribution
    requires, and the contacts (technical questions and pull requests:
    somol@utia.cas.cz, with "FST" in the subject line).
========================================================================== */
 
#include <memory>
#include <cassert>
#include <iostream>
#include <list>
#include "error.hpp"
#include "global.hpp"
#include "result_tracker_dupless.hpp"
 
/*============== Template parameter type naming conventions ==============
--------- Numeric types: -------------------------------------------------
DATATYPE - data sample values - usually real numbers (but may be integers
          in text processing etc.)
REALTYPE - must be real numbers - for representing intermediate results of 
          calculations like mean, covariance etc.
IDXTYPE - index values for enumeration of data samples - (nonnegative) integers, 
          extent depends on numbers of samples in data
DIMTYPE - index values for enumeration of features (dimensions), or classes (not 
          class sizes) - (nonnegative) integers, usually lower extent than IDXTYPE, 
          but be aware of expressions like _classes*_features*_features ! 
          in linearized representations of feature matrices for all classes
BINTYPE - feature selection marker type - represents ca. <10 different feature 
          states (selected, deselected, sel./desel. temporarily 1st nested loop, 2nd...)
RETURNTYPE - criterion value: real value, but may be extended in future to support 
          multiple values 
--------- Class types: ---------------------------------------------------
SUBSET       - class of class type Subset 
CLASSIFIER   - class implementing interface defined in abstract class Classifier 
EVALUATOR    - class implementing interface defined in abstract class Sequential_Step 
DISTANCE     - class implementing interface defined in abstract class Distance 
DATAACCESSOR - class implementing interface defined in abstract class Data_Accessor 
INTERVALCONTAINER - class of class type Data_Intervaller 
CONTAINER    - STL container of class type Data_Interval  
========================================================================== */
 
namespace FST {
 
template<class RETURNTYPE, typename IDXTYPE, typename DIMTYPE, class SUBSET, class CRITERION>
class Result_Tracker_Regularizer : public Clonable_Impl<Result_Tracker_Regularizer<RETURNTYPE,IDXTYPE,DIMTYPE,SUBSET,CRITERION>,Result_Tracker_Dupless<RETURNTYPE,IDXTYPE,DIMTYPE,SUBSET>,CLONE_INHERIT,CLONE_INHERIT,CLONE_DEEP> {
public:
        typedef Result_Tracker_Dupless<RETURNTYPE,IDXTYPE,DIMTYPE,SUBSET> parent; 
        typedef Clonable_Impl<Result_Tracker_Regularizer,parent,CLONE_INHERIT,CLONE_INHERIT,CLONE_DEEP> parent_impl; 
        typedef typename parent::ResultRec ResultRec; 
        typedef ResultRec* PResultRec; 
        typedef std::shared_ptr<SUBSET> PSubset; 
        typedef std::shared_ptr<CRITERION> PCriterion; 
        Result_Tracker_Regularizer(const IDXTYPE capacity_limit=0):parent_impl(capacity_limit),_searching(false) {notify("Result_Tracker_Regularizer constructor.");} 
        Result_Tracker_Regularizer(const Result_Tracker_Regularizer& rtr):parent_impl(rtr),_searching(false) {notify("Result_Tracker_Regularizer copy-constructor.");} 
        virtual ~Result_Tracker_Regularizer() {notify("Result_Tracker_Regularizer destructor.");}
        
        // NOTE: crit is expected to be already initialize()d before use here
        virtual bool optimize_within_margin(const RETURNTYPE margin, RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub, const PCriterion crit);

        virtual bool optimize_within_first_equivalence_group(RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub, const PCriterion crit);
        virtual bool optimize_within_next_equivalence_group(RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub);
 
 
 
        virtual std::ostream& print(std::ostream& os) const override {os << "Result_Tracker_Regularizer(limit=" << parent::_capacity_limit << ", margin=" << parent::_margin << ") size " << parent::results.size(); return os;}
 
private:
        bool optimize_within_equivalence_group(RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub); 
        PCriterion _crit; 
        RETURNTYPE _prevmax; 
        typename parent::CONSTRESULTSITER _iter; 
        bool _searching; 
};
 
 
 
template<class RETURNTYPE, typename IDXTYPE, typename DIMTYPE, class SUBSET, class CRITERION>
bool Result_Tracker_Regularizer<RETURNTYPE, IDXTYPE, DIMTYPE, SUBSET, CRITERION>::optimize_within_margin(const RETURNTYPE margin, RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub, const PCriterion crit)
{
        assert(margin>=0);
        if(parent::results.empty()) return false;
        bool stop=false;
        
        typename parent::CONSTRESULTSITER iter=parent::results.begin();
        typename parent::CONSTRESULTSITER sel=iter;
        RETURNTYPE bestval, val;
        
        // NOTE: crit is expected to be already initialize()d before use here
        if(!crit->evaluate(bestval,sel->sub)) return false;
        notify("sel: val1=",sel->value,", val2=",bestval,", getd ",sel->sub->get_d(),".");
        
        const RETURNTYPE threshold=sel->value-margin;
        iter++;
        while(!stop && iter!=parent::results.end())
        {
                if(iter->value<threshold) stop=true; else
                {
                        // NOTE: crit is expected to be already initialize()d before use here
                        if(!crit->evaluate(val,iter->sub)) return false;
                        notify("iter: val1=",iter->value,", val2=",val,", getd ",iter->sub->get_d(),".");
                        if(val>bestval || (val==bestval && sel->sub->get_d()>iter->sub->get_d())) {bestval=val; sel=iter; notify("best updated.");}
                        iter++;
                }
        }
        result1=sel->value; result2=bestval; sub=sel->sub;
        return true;
}
 
template<class RETURNTYPE, typename IDXTYPE, typename DIMTYPE, class SUBSET, class CRITERION>
bool Result_Tracker_Regularizer<RETURNTYPE, IDXTYPE, DIMTYPE, SUBSET, CRITERION>::optimize_within_equivalence_group(RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub)
{
        // precondition: _iter points to a valid record of the tracked list and _crit is set - both public entry points guarantee this before calling in
        _prevmax=_iter->value;
        RETURNTYPE val, maxval1, maxval2; PSubset maxsub=_iter->sub;
        maxval1=_iter->value;
        if(!_crit->evaluate(maxval2,_iter->sub)) {_searching=false; return false;}
        _iter++;
        while(_iter!=parent::results.end() && _iter->value==_prevmax)
        {
                if(!_crit->evaluate(val,_iter->sub)) {_searching=false; return false;}
                if(val>maxval2 || (val==maxval2 && maxsub->get_d()>_iter->sub->get_d())) {maxval2=val; maxsub=_iter->sub;}
                _iter++;
        }
        if(_iter==parent::results.end()) _searching=false; else _searching=true;
        result1=maxval1; result2=maxval2; sub=maxsub;
        return true;
}
 
template<class RETURNTYPE, typename IDXTYPE, typename DIMTYPE, class SUBSET, class CRITERION>
bool Result_Tracker_Regularizer<RETURNTYPE, IDXTYPE, DIMTYPE, SUBSET, CRITERION>::optimize_within_first_equivalence_group(RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub, const PCriterion crit)
{ // returns the subset from the first equivalence group (=subsets with equal primary criterion value) yielding the highest crit value
        assert(crit);
        assert(sub);
        if(parent::results.empty()) {_searching=false; return false;}
        _crit=crit;
        _iter=parent::results.begin();
        return optimize_within_equivalence_group(result1,result2,sub);
}
 
template<class RETURNTYPE, typename IDXTYPE, typename DIMTYPE, class SUBSET, class CRITERION>
bool Result_Tracker_Regularizer<RETURNTYPE, IDXTYPE, DIMTYPE, SUBSET, CRITERION>::optimize_within_next_equivalence_group(RETURNTYPE &result1, RETURNTYPE &result2, PSubset &sub) 
{ // returns the subset from the next equivalence group (=subsets with equal primary criterion value which is the next highest in sequence) yielding the highest crit value
        assert(_crit);
        assert(sub);
        if(!_searching) return false;
        return optimize_within_equivalence_group(result1,result2,sub);
}
 
} // namespace
#endif // FSTRESULTTRACKERREGULARIZER_H ///:~