FANNOptions Class Reference

Options for the FANN MVA method. More...

#include <FANN.h>

Inheritance diagram for FANNOptions:

Collaboration diagram for FANNOptions:

Public Member Functions
virtual void	load (const boost::property_tree::ptree &pt) override
	Load mechanism to load Options from a xml tree. More...
virtual void	save (boost::property_tree::ptree &pt) const override
	Save mechanism to store Options in a xml tree. More...
virtual po::options_description	getDescription () override
	Returns a program options description for all available options.
virtual std::string	getMethod () const override
	Return method name.
std::vector< unsigned int >	getHiddenLayerNeurons (unsigned int nf) const
	Returns the internal vector parameter with the number of hidden neurons per layer. More...

Public Attributes
unsigned int	m_max_epochs = 10000
	Maximum number of epochs.
bool	m_verbose_mode = true
	Sets to report training status or not.
std::string	m_hidden_layers_architecture = "3*N"
	String containing the architecture of hidden neurons. More...
std::string	m_hidden_activiation_function = "FANN_SIGMOID_SYMMETRIC"
	Activation function in hidden layer.
std::string	m_output_activiation_function = "FANN_SIGMOID_SYMMETRIC"
	Activation function in output layer.
std::string	m_error_function = "FANN_ERRORFUNC_LINEAR"
	Loss function.
std::string	m_training_method = "FANN_TRAIN_RPROP"
	Training method for back propagation.
double	m_validation_fraction = 0.5
	Fraction of training sample used for validation in order to avoid overtraining.
unsigned int	m_random_seeds
	Number of times the training is repeated with a new weight random seed. More...
unsigned int	m_test_rate
	Error on validation is compared with the one before. More...
unsigned int	m_number_of_threads = 8
	Number of threads for parallel training.
bool	m_scale_features = true
	Scale features before training.
bool	m_scale_target = true
	Scale target before training.

Detailed Description

Options for the FANN MVA method.

Definition at line 29 of file FANN.h.

Member Function Documentation

getHiddenLayerNeurons()

std::vector< unsigned int > getHiddenLayerNeurons

(

unsigned int

nf

)

const

Returns the internal vector parameter with the number of hidden neurons per layer.

Parameters

nf	number of features (input nodes).

Definition at line 93 of file FANNOptions.cc.

    {
      std::vector<unsigned int> hiddenLayers;
      std::stringstream iLayers(m_hidden_layers_architecture);
      std::string layer;
      while (std::getline(iLayers, layer, ',')) {
        for (auto& character : layer) {
          if (character == 'N') character = 'x';
        }
        auto* iLayerSize = new TFormula("iLayerSize", layer.c_str());
        hiddenLayers.push_back(iLayerSize->Eval(nf));
      }
      return hiddenLayers;
    }

load()

void load ( const boost::property_tree::ptree &  pt )

overridevirtual

Load mechanism to load Options from a xml tree.

Parameters

pt

xml tree

Implements Options.

Definition at line 21 of file FANNOptions.cc.

save()

void save ( boost::property_tree::ptree &  pt ) const

overridevirtual

Save mechanism to store Options in a xml tree.

Parameters

pt

xml tree

Implements Options.

Definition at line 48 of file FANNOptions.cc.

Member Data Documentation

m_hidden_layers_architecture

std::string m_hidden_layers_architecture = "3*N"

String containing the architecture of hidden neurons.

Ex. "3,3*N,3*(N-1)

Definition at line 62 of file FANN.h.

m_random_seeds

unsigned int m_random_seeds

Initial value:

=
        3

Number of times the training is repeated with a new weight random seed.

The one with the best result is saved.

Definition at line 70 of file FANN.h.

m_test_rate

unsigned int m_test_rate

Initial value:

=
        500

Error on validation is compared with the one before.

The number of epochs before is given by this parameter.

Definition at line 72 of file FANN.h.

---

The documentation for this class was generated from the following files:

• mva/methods/include/FANN.h
• mva/methods/src/FANNOptions.cc