Logo ROOT  
Reference Guide
 
Loading...
Searching...
No Matches
TMVA::ROCCurve Class Reference

Definition at line 46 of file ROCCurve.h.

Public Member Functions

 ROCCurve (const std::vector< Float_t > &mvaSignal, const std::vector< Float_t > &mvaBackground)
 
 ROCCurve (const std::vector< Float_t > &mvaSignal, const std::vector< Float_t > &mvaBackground, const std::vector< Float_t > &mvaSignalWeights, const std::vector< Float_t > &mvaBackgroundWeights)
 
 ROCCurve (const std::vector< Float_t > &mvaValues, const std::vector< Bool_t > &mvaTargets)
 
 ROCCurve (const std::vector< Float_t > &mvaValues, const std::vector< Bool_t > &mvaTargets, const std::vector< Float_t > &mvaWeights)
 
 ROCCurve (const std::vector< std::tuple< Float_t, Float_t, Bool_t > > &mvas)
 
 ~ROCCurve ()
 destructor
 
Double_t GetEffSForEffB (Double_t effB, const UInt_t num_points=41)
 Calculate the signal efficiency (sensitivity) for a given background efficiency (sensitivity).
 
const std::vector< std::tuple< Float_t, Float_t, Bool_t > > GetMvas () const
 
TGraphGetROCCurve (const UInt_t points=100)
 Returns a new TGraph containing the ROC curve.
 
Double_t GetROCIntegral (const UInt_t points=41)
 Calculates the ROC integral (AUC)
 

Private Member Functions

std::vector< Double_tComputeSensitivity (const UInt_t num_points)
 
std::vector< Double_tComputeSpecificity (const UInt_t num_points)
 
MsgLoggerLog () const
 

Private Attributes

TGraphfGraph
 
MsgLoggerfLogger
 ! message logger
 
std::vector< std::tuple< Float_t, Float_t, Bool_t > > fMva
 

#include <TMVA/ROCCurve.h>

Constructor & Destructor Documentation

◆ ROCCurve() [1/5]

TMVA::ROCCurve::ROCCurve ( const std::vector< std::tuple< Float_t, Float_t, Bool_t > > &  mvas)

Definition at line 47 of file ROCCurve.cxx.

◆ ROCCurve() [2/5]

TMVA::ROCCurve::ROCCurve ( const std::vector< Float_t > &  mvaValues,
const std::vector< Bool_t > &  mvaTargets,
const std::vector< Float_t > &  mvaWeights 
)

Definition at line 55 of file ROCCurve.cxx.

◆ ROCCurve() [3/5]

TMVA::ROCCurve::ROCCurve ( const std::vector< Float_t > &  mvaValues,
const std::vector< Bool_t > &  mvaTargets 
)

Definition at line 72 of file ROCCurve.cxx.

◆ ROCCurve() [4/5]

TMVA::ROCCurve::ROCCurve ( const std::vector< Float_t > &  mvaSignal,
const std::vector< Float_t > &  mvaBackground,
const std::vector< Float_t > &  mvaSignalWeights,
const std::vector< Float_t > &  mvaBackgroundWeights 
)

Definition at line 104 of file ROCCurve.cxx.

◆ ROCCurve() [5/5]

TMVA::ROCCurve::ROCCurve ( const std::vector< Float_t > &  mvaSignal,
const std::vector< Float_t > &  mvaBackground 
)

Definition at line 87 of file ROCCurve.cxx.

◆ ~ROCCurve()

TMVA::ROCCurve::~ROCCurve ( )

destructor

Definition at line 125 of file ROCCurve.cxx.

Member Function Documentation

◆ ComputeSensitivity()

std::vector< Double_t > TMVA::ROCCurve::ComputeSensitivity ( const UInt_t  num_points)
private

Definition at line 176 of file ROCCurve.cxx.

◆ ComputeSpecificity()

std::vector< Double_t > TMVA::ROCCurve::ComputeSpecificity ( const UInt_t  num_points)
private

Definition at line 140 of file ROCCurve.cxx.

◆ GetEffSForEffB()

Double_t TMVA::ROCCurve::GetEffSForEffB ( Double_t  effB,
const UInt_t  num_points = 41 
)

Calculate the signal efficiency (sensitivity) for a given background efficiency (sensitivity).

Parameters
effBBackground efficiency for which to calculate signal efficiency.
num_pointsNumber of points used for the underlying histogram. The number of bins will be num_points - 1.

Definition at line 219 of file ROCCurve.cxx.

◆ GetMvas()

const std::vector< std::tuple< Float_t, Float_t, Bool_t > > TMVA::ROCCurve::GetMvas ( ) const
inline

Definition at line 68 of file ROCCurve.h.

◆ GetROCCurve()

TGraph * TMVA::ROCCurve::GetROCCurve ( const UInt_t  num_points = 100)

Returns a new TGraph containing the ROC curve.

Sensitivity is on the x-axis, specificity on the y-axis.

Parameters
num_pointsGranularity of the resulting curve. The curve will be subdivided into num_points - 1 regions where the performance of the classifier is sampled. Larger number means more accurate, but more costly, evaluation.

Definition at line 276 of file ROCCurve.cxx.

◆ GetROCIntegral()

Double_t TMVA::ROCCurve::GetROCIntegral ( const UInt_t  num_points = 41)

Calculates the ROC integral (AUC)

Parameters
num_pointsGranularity of the resulting curve used for integration. The curve will be subdivided into num_points - 1 regions where the performance of the classifier is sampled. Larger number means more accurate, but more costly, evaluation.

Definition at line 250 of file ROCCurve.cxx.

◆ Log()

TMVA::MsgLogger & TMVA::ROCCurve::Log ( ) const
private

Definition at line 130 of file ROCCurve.cxx.

Member Data Documentation

◆ fGraph

TGraph* TMVA::ROCCurve::fGraph
private

Definition at line 73 of file ROCCurve.h.

◆ fLogger

MsgLogger* TMVA::ROCCurve::fLogger
mutableprivate

! message logger

Definition at line 70 of file ROCCurve.h.

◆ fMva

std::vector<std::tuple<Float_t, Float_t, Bool_t> > TMVA::ROCCurve::fMva
private

Definition at line 75 of file ROCCurve.h.

Libraries for TMVA::ROCCurve:

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