example_perceptron_kernel_poly_band.cpp

Programa para el testeo de las funciones.Para compilar o código example_perceptron_kernel_poly_band.cpp:

g++ -static -o example_perceptron_kernel_poly_band example_perceptron_kernel_poly_band.cpp -lpdsnnmm 

Para executar o programa:

./example_perceptron_kernel_poly_band

Retornando por consola:

-------------------------
   IterationConf data    
-------------------------
    Alpha: 0.1
   Lambda: 0.1
 MinError: 1e-05
  MaxIter: 1000
-------------------------
LastError: 1.9828e-05
 LastIter: 1000
    
Neurona:
-4.388764100926 0.46547441220239        0.024070838644976       13.586974277052 -0.48056582017806       13.411529224111
    
-----------------------------
      ClassificationMetrics data      
-----------------------------
Threshold       : 0.31
Samples         : 2000
----------------:------------
Pred.[0]/Act.[*]: 999   10
Pred.[1]/Act.[*]: 1     990
----------------:------------
Accuracy        : 99.45%
Precision       : 99.9%
Recall          : 99%
----------------:------------
FScore          : 99.45%
    
-----------------------------
      ClassificationMetrics data      
-----------------------------
Threshold       : 0.31
Samples         : 2000
----------------:------------
Pred.[0]/Act.[*]: 997   14
Pred.[1]/Act.[*]: 3     986
----------------:------------
Accuracy        : 99.15%
Precision       : 99.7%
Recall          : 98.6%
----------------:------------
FScore          : 99.15%

Código example_perceptron_kernel_poly_band.cpp:

 
#include <iostream>
#include <Pds/Ra>
#include <Pds/Ml>
 
    
int main(void)
{
    Pds::ClassificationMetrics Metrics;
    Pds::Vector Yp;
    Pds::Matrix F;
    Pds::IterationConf Conf;
    unsigned int M=4;
    
    // Generating data
    unsigned int L=1000;
    Pds::Matrix X;
    Pds::Vector Y;
    Pds::DataSet::Classification::LoadDataBand2(L,X,Y);
        
    // Create Perceptron
    F=Pds::Kernel::Polynomial(X,M);
    Conf.SetMaxIter(10000);
    Pds::Perceptron Neurona(Conf,F,Y);
    Neurona.Print("\nNeurona:\n");
    
    // Evaluate training data
    Yp=Neurona.Evaluate(F);
    
    // Metrics of training
    Metrics = Pds::ClassificationMetrics::Optimal(Yp,Y);
    Metrics.Print("\n");
        
    // Evaluate testing data
    Pds::DataSet::Classification::LoadDataBand2(L,X,Y); // Load new data
    F=Pds::Kernel::Polynomial(X,M);
    Yp=Neurona.Evaluate(F);
    
    // Metrics testing
    Metrics = Pds::ClassificationMetrics::Calculate(Metrics.Threshold,Yp,Y);
    Metrics.Print("\n");
    return 0;
}