example_decisiontree.cpp

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

g++ -static -o example_decisiontree example_decisiontree.cpp -lpdsmlmm 

Para executar o programa:

./example_decisiontree

Retornando por consola:

Pds::DecisionTree::Counter: 123

Training metrics:
╔═══════════════════════════════════╗
║    ClassificationMetrics data     ║
╠═══════════════════════════════════╣
║        Threshold: 0.5             ║
║          Samples: 2000            ║
╠═══════════════════════════════════╣
║ Pred.[0]/Act.[*]: 982     3       ║
║ Pred.[1]/Act.[*]: 18      997     ║
╠═══════════════════════════════════╣
║         Accuracy: 98.95 %         ║
║        Precision: 98.23 %         ║
║           Recall: 99.7  %         ║
╠═══════════════════════════════════╣
║           FScore: 98.96 %         ║
╚═══════════════════════════════════╝

Testing metrics:
╔═══════════════════════════════════╗
║    ClassificationMetrics data     ║
╠═══════════════════════════════════╣
║        Threshold: 0.5             ║
║          Samples: 2000            ║
╠═══════════════════════════════════╣
║ Pred.[0]/Act.[*]: 955     41      ║
║ Pred.[1]/Act.[*]: 45      959     ║
╠═══════════════════════════════════╣
║         Accuracy: 95.7  %         ║
║        Precision: 95.52 %         ║
║           Recall: 95.9  %         ║
╠═══════════════════════════════════╣
║           FScore: 95.71 %         ║
╚═══════════════════════════════════╝

Clasification - data

Clasification - training

Clasification - testing

Código example_decisiontree.cpp:

 
#include <iostream>
#include <Pds/Ra>
#include <Pds/Ml>
    
int main(void)
{
    Pds::Octave::Colormap="jet";
    Pds::Octave::XLabel="x_1";
    Pds::Octave::YLabel="x_2";
    
    Pds::ClassificationMetrics Metrics;
    
    Pds::IterationConf Conf;
    Conf.Show=false; Conf.SetMinError(1e-07);    Conf.SetAlpha(0.1);
    
    Pds::Vector Yp;
    
 
    // Load training data
    unsigned int L=1000;
    Pds::Matrix X;
    Pds::Vector Y;
    Pds::DataSet::Classification::LoadDataYinYang(L,X,Y);
    
    Pds::Octave::Plot::ScatterX2DY(X,Y,"testando.m","example_decisiontree_data.png");
    
    // Creando un arbol
    Pds::DecisionTree::Counter=0;
    Pds::DecisionTree Arbol(Conf,X,Y,0.8,4);
    std::cout<<"Pds::DecisionTree::Counter: "<<Pds::DecisionTree::Counter<<std::endl;
 
    // Exportando arbol en archivo en formato dot.
    Arbol.ExportDotFile("arbol.dot");
    
    // Evaluate training data
    Yp=Arbol.Evaluate(X);
    
    // Metrics of training
    Metrics = Pds::ClassificationMetrics::Calculate(0.5,Yp,Y);
    Metrics.Print("\nTraining metrics:\n");
    
    Pds::Octave::Plot::ScatterX2DY(X,Yp,"testando.m","example_decisiontree_training.png");
    
    // Load testing data
    Pds::DataSet::Classification::LoadDataYinYang(L,X,Y); // Load new data
    
    // Evaluate testing data
    Yp=Arbol.Evaluate(X);
    
    // Metrics of testing
    Metrics = Pds::ClassificationMetrics::Calculate(Metrics.Threshold,Yp,Y);
    Metrics.Print("\nTesting metrics:\n");
    
    Pds::Octave::Plot::ScatterX2DY(X,Yp,"testando.m","example_decisiontree_testing.png");
 
    return 0;
}