scaling.h

//   OpenNN: Open Neural Networks Library
//   www.opennn.net
//
//   S T A T I S T I C S   H E A D E R
//
//   Artificial Intelligence Techniques, SL
//   artelnics@artelnics.com
 
#ifndef SCALING_H
#define SCALING_H
 
// System includes
 
#include <fstream>
#include <iostream>
#include <limits>
#include <math.h>
#include <vector>
 
// OpenNN includes
 
#include "config.h"
#include "statistics.h"
 
using namespace std;
using namespace Eigen;
 
namespace OpenNN
{
 
    enum class Scaler{NoScaling, MinimumMaximum, MeanStandardDeviation, StandardDeviation, Logarithm};
 
    void scale_mean_standard_deviation(Tensor<type, 2>&, const Index&, const Descriptives&);
    void scale_standard_deviation(Tensor<type, 2>&, const Index&, const Descriptives&);
    void scale_minimum_maximum(Tensor<type, 2>&, const Index&, const Descriptives&, const type& = type(-1), const type& = type(1));
 
    Tensor<type, 1> scale_minimum_maximum(const Tensor<type, 1>&);
    Tensor<type, 2> scale_minimum_maximum(const Tensor<type, 2>&);
 
    void scale_logarithmic(Tensor<type, 2>&, const Index&);
    void scale_minimum_maximum_binary(Tensor<type, 2>&, const type&, const type&, const Index&);
 
    void unscale_minimum_maximum(Tensor<type, 2>&, const Index&, const Descriptives&, const type& = type(-1), const type& = type(1));
    void unscale_mean_standard_deviation(Tensor<type, 2>&, const Index&, const Descriptives&);
    void unscale_standard_deviation(Tensor<type, 2>&, const Index&, const Descriptives&);
    void unscale_logarithmic(Tensor<type, 2>&, const Index&);
 
}
 
#endif // STATISTICS_H