Wraps another Dataset and provides a view to a subset of its features and events. More...

#include <Dataset.h>

Inheritance diagram for SubDataset:

Public Member Functions
	SubDataset (const GeneralOptions &general_options, const std::vector< bool > &events, Dataset &dataset)
	Constructs a new SubDataset holding a reference to the wrapped Dataset.

virtual unsigned int	getNumberOfFeatures () const override
	Returns the number of features in this dataset, so the size of the given subset of the variables.

virtual unsigned int	getNumberOfSpectators () const override
	Returns the number of spectators in this dataset, so the size of the given subset of the spectators.

virtual unsigned int	getNumberOfEvents () const override
	Returns the number of events in the wrapped dataset.

virtual void	loadEvent (unsigned int iEvent) override
	Load the event number iEvent from the wrapped dataset.

virtual std::vector< float >	getFeature (unsigned int iFeature) override
	Returns all values of one feature in a std::vector<float> of the wrapped dataset.

virtual std::vector< float >	getSpectator (unsigned int iSpectator) override
	Returns all values of one spectator in a std::vector<float> of the wrapped dataset.

virtual float	getSignalFraction ()
	Returns the signal fraction of the whole sample.

virtual unsigned int	getFeatureIndex (const std::string &feature)
	Return index of feature with the given name.

virtual unsigned int	getSpectatorIndex (const std::string &spectator)
	Return index of spectator with the given name.

virtual std::vector< float >	getWeights ()
	Returns all weights.

virtual std::vector< float >	getTargets ()
	Returns all targets.

virtual std::vector< bool >	getSignals ()
	Returns all is Signals.

Public Attributes
GeneralOptions	m_general_options
	GeneralOptions passed to this dataset.

std::vector< float >	m_input
	Contains all feature values of the currently loaded event.

std::vector< float >	m_spectators
	Contains all spectators values of the currently loaded event.

float	m_weight
	Contains the weight of the currently loaded event.

float	m_target
	Contains the target value of the currently loaded event.

bool	m_isSignal
	Defines if the currently loaded event is signal or background.

Private Attributes
bool	m_use_event_indices = false
	Use only a subset of the wrapped dataset events.

std::vector< unsigned int >	m_feature_indices
	Mapping from the position of a feature in the given subset to its position in the wrapped dataset.

std::vector< unsigned int >	m_spectator_indices
	Mapping from the position of a spectator in the given subset to its position in the wrapped dataset.

std::vector< unsigned int >	m_event_indices
	Mapping from the position of a event in the given subset to its position in the wrapped dataset.

Dataset &	m_dataset
	Reference to the wrapped dataset.

Detailed Description

Wraps another Dataset and provides a view to a subset of its features and events.

Used by the Combination method which can combine multiple methods with possibly different variables

Definition at line 234 of file Dataset.h.

Constructor & Destructor Documentation

◆ SubDataset()

SubDataset	(	const GeneralOptions &	general_options,
		const std::vector< bool > &	events,
		Dataset &	dataset
	)

Constructs a new SubDataset holding a reference to the wrapped Dataset.

Parameters

general_options	which defines e.g. a subset of variables of the original dataset
events	subset of events which are provided by this Dataset
dataset	reference to the wrapped Dataset

Definition at line 186 of file Dataset.cc.

                                             : Dataset(general_options), m_dataset(dataset)
    {
 
      for (auto& v : m_general_options.m_variables) {
        auto it = std::find(m_dataset.m_general_options.m_variables.begin(), m_dataset.m_general_options.m_variables.end(), v);
        if (it == m_dataset.m_general_options.m_variables.end()) {
          B2ERROR("Couldn't find variable " << v << " in GeneralOptions");
          throw std::runtime_error("Couldn't find variable " + v + " in GeneralOptions");
        }
        m_feature_indices.push_back(it - m_dataset.m_general_options.m_variables.begin());
      }
 
      for (auto& v : m_general_options.m_spectators) {
        auto it = std::find(m_dataset.m_general_options.m_spectators.begin(), m_dataset.m_general_options.m_spectators.end(), v);
        if (it == m_dataset.m_general_options.m_spectators.end()) {
          B2ERROR("Couldn't find spectator " << v << " in GeneralOptions");
          throw std::runtime_error("Couldn't find spectator " + v + " in GeneralOptions");
        }
        m_spectator_indices.push_back(it - m_dataset.m_general_options.m_spectators.begin());
      }
 
      if (events.size() > 0)
        m_use_event_indices = true;
 
      if (m_use_event_indices) {
        m_event_indices.resize(dataset.getNumberOfEvents());
        unsigned int n_events = 0;
        for (unsigned int iEvent = 0; iEvent < dataset.getNumberOfEvents(); ++iEvent) {
          if (events.size() == 0 or events[iEvent]) {
            m_event_indices[n_events] = iEvent;
            n_events++;
          }
        }
        m_event_indices.resize(n_events);
      }
 
    }

Member Function Documentation

◆ getFeature()

std::vector< float > getFeature ( unsigned int iFeature )

overridevirtual

Returns all values of one feature in a std::vector<float> of the wrapped dataset.

Parameters

iFeature the position of the feature to return in the given subset

Reimplemented from Dataset.

Definition at line 245 of file Dataset.cc.

    {
 
      auto v = m_dataset.getFeature(m_feature_indices[iFeature]);
      if (not m_use_event_indices)
        return v;
      std::vector<float> result(m_event_indices.size());
      for (unsigned int iEvent = 0; iEvent < getNumberOfEvents(); ++iEvent) {
        result[iEvent] = v[m_event_indices[iEvent]];
      }
      return result;
 
    }

◆ getFeatureIndex()

unsigned int getFeatureIndex ( const std::string & feature )

virtualinherited

Return index of feature with the given name.

Parameters

feature name of the feature

Definition at line 50 of file Dataset.cc.

    {
 
      auto it = std::find(m_general_options.m_variables.begin(), m_general_options.m_variables.end(), feature);
      if (it == m_general_options.m_variables.end()) {
        B2ERROR("Unknown feature named " << feature);
        return 0;
      }
      return std::distance(m_general_options.m_variables.begin(), it);
 
    }

◆ getNumberOfEvents()

virtual unsigned int getNumberOfEvents ( ) const

inlineoverridevirtual

Returns the number of events in the wrapped dataset.

Implements Dataset.

Definition at line 258 of file Dataset.h.

258{ return m_use_event_indices ? m_event_indices.size() : m_dataset.getNumberOfEvents(); }

Belle2::MVA::Dataset::getNumberOfEvents

virtual unsigned int getNumberOfEvents() const =0

Returns the number of events in this dataset.

◆ getNumberOfFeatures()

virtual unsigned int getNumberOfFeatures ( ) const

inlineoverridevirtual

Returns the number of features in this dataset, so the size of the given subset of the variables.

Implements Dataset.

Definition at line 248 of file Dataset.h.

248{ return m_feature_indices.size(); }

◆ getNumberOfSpectators()

virtual unsigned int getNumberOfSpectators ( ) const

inlineoverridevirtual

Returns the number of spectators in this dataset, so the size of the given subset of the spectators.

Implements Dataset.

Definition at line 253 of file Dataset.h.

253{ return m_spectator_indices.size(); }

◆ getSignalFraction()

float getSignalFraction ( )

virtualinherited

Returns the signal fraction of the whole sample.

Reimplemented in SPlotDataset.

Definition at line 35 of file Dataset.cc.

    {
 
      double signal_weight_sum = 0;
      double weight_sum = 0;
      for (unsigned int i = 0; i < getNumberOfEvents(); ++i) {
        loadEvent(i);
        weight_sum += m_weight;
        if (m_isSignal)
          signal_weight_sum += m_weight;
      }
      return signal_weight_sum / weight_sum;
 
    }

◆ getSignals()

std::vector< bool > getSignals ( )

virtualinherited

Returns all is Signals.

Reimplemented in ReweightingDataset.

Definition at line 122 of file Dataset.cc.

    {
 
      std::vector<bool> result(getNumberOfEvents());
      for (unsigned int iEvent = 0; iEvent < getNumberOfEvents(); ++iEvent) {
        loadEvent(iEvent);
        result[iEvent] = m_isSignal;
      }
      return result;
 
    }

◆ getSpectator()

std::vector< float > getSpectator ( unsigned int iSpectator )

overridevirtual

Returns all values of one spectator in a std::vector<float> of the wrapped dataset.

Parameters

iSpectator the position of the spectator to return in the given subset

Reimplemented from Dataset.

Definition at line 259 of file Dataset.cc.

    {
 
      auto v = m_dataset.getSpectator(m_spectator_indices[iSpectator]);
      if (not m_use_event_indices)
        return v;
      std::vector<float> result(m_event_indices.size());
      for (unsigned int iEvent = 0; iEvent < getNumberOfEvents(); ++iEvent) {
        result[iEvent] = v[m_event_indices[iEvent]];
      }
      return result;
 
    }

◆ getSpectatorIndex()

unsigned int getSpectatorIndex ( const std::string & spectator )

virtualinherited

Return index of spectator with the given name.

Parameters

spectator name of the spectator

Definition at line 62 of file Dataset.cc.

    {
 
      auto it = std::find(m_general_options.m_spectators.begin(), m_general_options.m_spectators.end(), spectator);
      if (it == m_general_options.m_spectators.end()) {
        B2ERROR("Unknown spectator named " << spectator);
        return 0;
      }
      return std::distance(m_general_options.m_spectators.begin(), it);
 
    }

◆ getTargets()

std::vector< float > getTargets ( )

virtualinherited

Returns all targets.

Reimplemented in RegressionDataSet, and ReweightingDataset.

Definition at line 110 of file Dataset.cc.

    {
 
      std::vector<float> result(getNumberOfEvents());
      for (unsigned int iEvent = 0; iEvent < getNumberOfEvents(); ++iEvent) {
        loadEvent(iEvent);
        result[iEvent] = m_target;
      }
      return result;
 
    }

◆ getWeights()

std::vector< float > getWeights ( )

virtualinherited

Returns all weights.

Reimplemented in ROOTDataset, RegressionDataSet, and ReweightingDataset.

Definition at line 98 of file Dataset.cc.

    {
 
      std::vector<float> result(getNumberOfEvents());
      for (unsigned int iEvent = 0; iEvent < getNumberOfEvents(); ++iEvent) {
        loadEvent(iEvent);
        result[iEvent] = m_weight;
      }
      return result;
 
    }

◆ loadEvent()

void loadEvent ( unsigned int iEvent )

overridevirtual

Load the event number iEvent from the wrapped dataset.

Parameters

iEvent event number to load

Implements Dataset.

Definition at line 225 of file Dataset.cc.

    {
      unsigned int index = iEvent;
      if (m_use_event_indices)
        index = m_event_indices[iEvent];
      m_dataset.loadEvent(index);
      m_target = m_dataset.m_target;
      m_weight = m_dataset.m_weight;
      m_isSignal = m_dataset.m_isSignal;
 
      for (unsigned int iFeature = 0; iFeature < m_input.size(); ++iFeature) {
        m_input[iFeature] = m_dataset.m_input[m_feature_indices[iFeature]];
      }
 
      for (unsigned int iSpectator = 0; iSpectator < m_spectators.size(); ++iSpectator) {
        m_spectators[iSpectator] = m_dataset.m_spectators[m_spectator_indices[iSpectator]];
      }
 
    }

Definition at line 279 of file Dataset.h.

◆ m_weight

float m_weight

inherited

Contains the weight of the currently loaded event.

Definition at line 125 of file Dataset.h.

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

mva/interface/include/Dataset.h
mva/interface/src/Dataset.cc

Public Member Functions

Public Attributes

Private Attributes

Detailed Description

Constructor & Destructor Documentation

◆ SubDataset()

Member Function Documentation

◆ getFeature()

◆ getFeatureIndex()

◆ getNumberOfEvents()

◆ getNumberOfFeatures()

◆ getNumberOfSpectators()

◆ getSignalFraction()

◆ getSignals()

◆ getSpectator()

◆ getSpectatorIndex()

◆ getTargets()

◆ getWeights()

◆ loadEvent()

Member Data Documentation

◆ m_dataset

◆ m_event_indices

◆ m_feature_indices

◆ m_general_options

◆ m_input

◆ m_isSignal

◆ m_spectator_indices

◆ m_spectators

◆ m_target

◆ m_use_event_indices

◆ m_weight