data_scaler_white.hpp

Go to the documentation of this file.

#ifndef FSTDATASCALERWHITE_H
#define FSTDATASCALERWHITE_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 <iostream>
#include <cassert>
#include <sstream>
#include <cmath>
#include "error.hpp"
#include "global.hpp"
#include "data_scaler.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<typename DATATYPE, typename IDXTYPE>
class Data_Scaler_white : public Data_Scaler<DATATYPE> {
public:
        Data_Scaler_white(const int dims=1): learn_loop(0), count1(0), mean(0.0), count2(0), stddev(0.0), missing_values(false), _missing_val_code(0) {if(dims!=1) throw fst_error("Data_Scaler_white() implemented for 1-dimensional data only."); notify("Data_Scaler_white constructor.");} 
        Data_Scaler_white(const DATATYPE missing_val_code, const int dims): learn_loop(0), count1(0), mean(0.0), count2(0), stddev(0.0), missing_values(true), _missing_val_code(missing_val_code) {if(dims!=1) throw fst_error("Data_Scaler_white() implemented for 1-dimensional data only."); notify("Data_Scaler_white constructor.");} 
        Data_Scaler_white(const Data_Scaler_white &ds) : learn_loop(ds.learn_loop), count1(ds.count1), mean(ds.mean), count2(ds.count2), stddev(ds.stddev), missing_values(ds.missing_values), _missing_val_code(ds._missing_val_code)  {notify("Data_Scaler_white copy constructor.");} 
        virtual ~Data_Scaler_white() {notify("Data_Scaler_white destructor.");}
        virtual int learn_loops() const override {return 2;} 
        virtual bool startFirstLoop() override; 
        virtual bool startNextLoop() override; 
        virtual void learn(const DATATYPE& value) override; 
 
        virtual DATATYPE scale(const DATATYPE& value) override; 
        virtual void scale_inplace(DATATYPE& value) override; 
 
        virtual Data_Scaler_white* duplicate() const override {return new Data_Scaler_white(*this);}
 
        virtual std::ostream& print(std::ostream& os) const override {os << "Data_Scaler_white("; if(missing_values) os << "missing_value=" << _missing_val_code; os << ")"; return os;}
        
protected:
        int learn_loop; 
        IDXTYPE count1; 
        DATATYPE mean; 
        IDXTYPE count2; 
        DATATYPE stddev; 
 
        // missing value substitution support
        bool missing_values; 
        const DATATYPE _missing_val_code; 
};
 
template<typename DATATYPE, typename IDXTYPE>
bool Data_Scaler_white<DATATYPE,IDXTYPE>::startFirstLoop() 
{
        learn_loop=1; count1=0; mean=0.0;
        return true;
}
        
template<typename DATATYPE, typename IDXTYPE>
bool Data_Scaler_white<DATATYPE,IDXTYPE>::startNextLoop() 
{
        if(learn_loop==1) {
                 if(count1>0) mean/=count1; else mean=0.0;
                 learn_loop=2; count2=0; stddev=0.0; return true;
        }
        else if(learn_loop==2) {
                 // population variance: the sum of squared deviations is divided by the number of values,
                 // not by that number minus one
                 if(count2>0) stddev=sqrt(stddev/count2); else stddev=0.0; return false;
        }
        else {
                std::ostringstream sos; sos << "Data_Scaler_white::startNextLoop() wrong previous loop. (count1=" << count1 << ", count2=" << count2 << ")";
                throw fst_error(sos.str());
        }
}
 
template<typename DATATYPE, typename IDXTYPE>
void Data_Scaler_white<DATATYPE,IDXTYPE>::learn(const DATATYPE& value)
{
        if(learn_loop==1) {
                if(!missing_values || value!=_missing_val_code) {
                        ++count1;
                        mean+=value;
                }
        } else if(learn_loop==2) {
                if(!missing_values || value!=_missing_val_code) {
                        ++count2;
                        stddev+=(value-mean)*(value-mean);
                }
        } else throw fst_error("Data_Scaler_white::learn() wrong loop.");
}
 
template<typename DATATYPE, typename IDXTYPE>
DATATYPE Data_Scaler_white<DATATYPE,IDXTYPE>::scale(const DATATYPE& value) 
{
        assert(learn_loop==2 && count1==count2);
        if(missing_values && value==_missing_val_code) return 0; // substitute the mean, which in whitened coordinates is 0
        if(stddev>0.0) return (value-mean)/stddev; else return 0;
}
        
template<typename DATATYPE, typename IDXTYPE>
void Data_Scaler_white<DATATYPE,IDXTYPE>::scale_inplace(DATATYPE& value) 
{
        assert(learn_loop==2 && count1==count2);
        if(missing_values && value==_missing_val_code) {value=0; return;} // substitute the mean, which in whitened coordinates is 0
        if(stddev>0.0) value=(value-mean)/stddev; else value=0;
}
 
} // namespace
#endif // FSTDATASCALERWHITE_H ///:~