Logo ROOT  

Reference Guide
 
Loading...
Searching...
No Matches
List of all members | Public Member Functions | Private Member Functions | Private Attributes | List of all members
TMVA::ROCCalc Class Reference
TMVA

Definition at line 20 of file ROCCalc.h.

Public Member Functions

 ROCCalc (TH1 *mvaS, TH1 *mvaB)
 
 ~ROCCalc ()
 destructor
 
void ApplySignalAndBackgroundStyle (TH1 *sig, TH1 *bkg, TH1 *any=nullptr)
 
Double_t GetEffSForEffBof (Double_t effBref, Double_t &effSerr)
 get the signal efficiency for a particular background efficiency that will be the value of the efficiency retured (does not affect the efficiency-vs-bkg plot which is done anyway.
 
TH1 * GetMvaBpdf ()
 
TH1 * GetMvaSpdf ()
 
TH1 * GetPurity (Int_t nStot, Int_t nBtot)
 
TH1D * GetROC ()
 get the ROC curve
 
Double_t GetROCIntegral ()
 code to compute the area under the ROC ( rej-vs-eff ) curve
 
Double_t GetSignalReferenceCut ()
 
TH1 * GetSignificance (Int_t nStot, Int_t nBtot)
 
Bool_t GetStatus ()
 
void ResetStatus ()
 

Private Member Functions

Double_t GetEffForRoot (Double_t theCut)
 returns efficiency as function of cut
 
MsgLogger & Log () const
 message logger
 
Double_t Root (Double_t)
 Root finding using Brents algorithm; taken from CERNLIB function RZERO.
 

Private Attributes

TH1D * effBvsS
 
Double_t fAbsTol
 absolute tolerance deviation
 
Int_t fCutOrientation
 +1 if larger mva value means more signal like, -1 otherwise
 
MsgLogger * fLogger
 
Int_t fMaxIter
 maximum number of iterations
 
TH1 * fmvaB
 the input mva distributions
 
TH1 * fmvaBcumul
 
TH1 * fmvaBpdf
 the normalized (and rebinned) input mva distributions
 
TH1 * fmvaS
 
TH1 * fmvaScumul
 
TH1 * fmvaSpdf
 
UInt_t fNbins
 
Int_t fnBtot
 
Double_t fNevtS
 number of signal events (used in error calculation)
 
Int_t fnStot
 
TH1 * fPurity
 
Double_t fSignalCut
 MVA cut value for last demanded background rejection or signal efficiency.
 
TH1 * fSignificance
 
TSpline * fSplB
 
TSpline * fSpleffBvsS
 
TSpline * fSplmvaCumB
 spline of cumulated mva distributions
 
TSpline * fSplmvaCumS
 
TSpline * fSplS
 
Bool_t fStatus
 false if is found some error in mvaS or mvaB
 
Bool_t fUseSplines
 
Float_t fXmax
 min and max of the mva distribution
 
Float_t fXmin
 
TH1D * inveffBvsS
 
TH1D * rejBvsS
 

#include <TMVA/ROCCalc.h>

Constructor & Destructor Documentation

◆ ROCCalc()

TMVA::ROCCalc::ROCCalc ( TH1 *  mvaS,
TH1 *  mvaB 
)

Definition at line 42 of file ROCCalc.cxx.

◆ ~ROCCalc()

TMVA::ROCCalc::~ROCCalc ( )

destructor

Definition at line 160 of file ROCCalc.cxx.

Member Function Documentation

◆ ApplySignalAndBackgroundStyle()

void TMVA::ROCCalc::ApplySignalAndBackgroundStyle ( TH1 *  sig,
TH1 *  bkg,
TH1 *  any = nullptr 
)

Definition at line 113 of file ROCCalc.cxx.

◆ GetEffForRoot()

Double_t TMVA::ROCCalc::GetEffForRoot ( Double_t  theCut)
private

returns efficiency as function of cut

Definition at line 323 of file ROCCalc.cxx.

◆ GetEffSForEffBof()

Double_t TMVA::ROCCalc::GetEffSForEffBof ( Double_t  effBref,
Double_t &  effSerr 
)

get the signal efficiency for a particular background efficiency that will be the value of the efficiency retured (does not affect the efficiency-vs-bkg plot which is done anyway.

Definition at line 292 of file ROCCalc.cxx.

◆ GetMvaBpdf()

TH1 * TMVA::ROCCalc::GetMvaBpdf ( )
inline

Definition at line 42 of file ROCCalc.h.

◆ GetMvaSpdf()

TH1 * TMVA::ROCCalc::GetMvaSpdf ( )
inline

Definition at line 41 of file ROCCalc.h.

◆ GetPurity()

TH1 * TMVA::ROCCalc::GetPurity ( Int_t  nStot,
Int_t  nBtot 
)

Definition at line 429 of file ROCCalc.cxx.

◆ GetROC()

TH1D * TMVA::ROCCalc::GetROC ( )

get the ROC curve

Definition at line 178 of file ROCCalc.cxx.

◆ GetROCIntegral()

Double_t TMVA::ROCCalc::GetROCIntegral ( )

code to compute the area under the ROC ( rej-vs-eff ) curve

Definition at line 268 of file ROCCalc.cxx.

◆ GetSignalReferenceCut()

Double_t TMVA::ROCCalc::GetSignalReferenceCut ( )
inline

Definition at line 32 of file ROCCalc.h.

◆ GetSignificance()

TH1 * TMVA::ROCCalc::GetSignificance ( Int_t  nStot,
Int_t  nBtot 
)

Definition at line 440 of file ROCCalc.cxx.

◆ GetStatus()

Bool_t TMVA::ROCCalc::GetStatus ( )
inline

Definition at line 45 of file ROCCalc.h.

◆ Log()

MsgLogger & TMVA::ROCCalc::Log ( ) const
inlineprivate

message logger

Definition at line 78 of file ROCCalc.h.

◆ ResetStatus()

void TMVA::ROCCalc::ResetStatus ( )
inline

Definition at line 46 of file ROCCalc.h.

◆ Root()

Double_t TMVA::ROCCalc::Root ( Double_t  refValue)
private

Root finding using Brents algorithm; taken from CERNLIB function RZERO.

Definition at line 348 of file ROCCalc.cxx.

Member Data Documentation

◆ effBvsS

TH1D* TMVA::ROCCalc::effBvsS
private

Definition at line 71 of file ROCCalc.h.

◆ fAbsTol

Double_t TMVA::ROCCalc::fAbsTol
private

absolute tolerance deviation

Definition at line 52 of file ROCCalc.h.

◆ fCutOrientation

Int_t TMVA::ROCCalc::fCutOrientation
private

+1 if larger mva value means more signal like, -1 otherwise

Definition at line 63 of file ROCCalc.h.

◆ fLogger

MsgLogger* TMVA::ROCCalc::fLogger
mutableprivate

Definition at line 77 of file ROCCalc.h.

◆ fMaxIter

Int_t TMVA::ROCCalc::fMaxIter
private

maximum number of iterations

Definition at line 51 of file ROCCalc.h.

◆ fmvaB

TH1 * TMVA::ROCCalc::fmvaB
private

the input mva distributions

Definition at line 59 of file ROCCalc.h.

◆ fmvaBcumul

TH1 * TMVA::ROCCalc::fmvaBcumul
private

Definition at line 67 of file ROCCalc.h.

◆ fmvaBpdf

TH1 * TMVA::ROCCalc::fmvaBpdf
private

the normalized (and rebinned) input mva distributions

Definition at line 60 of file ROCCalc.h.

◆ fmvaS

TH1* TMVA::ROCCalc::fmvaS
private

Definition at line 59 of file ROCCalc.h.

◆ fmvaScumul

TH1* TMVA::ROCCalc::fmvaScumul
private

Definition at line 67 of file ROCCalc.h.

◆ fmvaSpdf

TH1* TMVA::ROCCalc::fmvaSpdf
private

Definition at line 60 of file ROCCalc.h.

◆ fNbins

UInt_t TMVA::ROCCalc::fNbins
private

Definition at line 56 of file ROCCalc.h.

◆ fnBtot

Int_t TMVA::ROCCalc::fnBtot
private

Definition at line 68 of file ROCCalc.h.

◆ fNevtS

Double_t TMVA::ROCCalc::fNevtS
private

number of signal events (used in error calculation)

Definition at line 62 of file ROCCalc.h.

◆ fnStot

Int_t TMVA::ROCCalc::fnStot
private

Definition at line 68 of file ROCCalc.h.

◆ fPurity

TH1* TMVA::ROCCalc::fPurity
private

Definition at line 70 of file ROCCalc.h.

◆ fSignalCut

Double_t TMVA::ROCCalc::fSignalCut
private

MVA cut value for last demanded background rejection or signal efficiency.

Definition at line 75 of file ROCCalc.h.

◆ fSignificance

TH1* TMVA::ROCCalc::fSignificance
private

Definition at line 69 of file ROCCalc.h.

◆ fSplB

TSpline * TMVA::ROCCalc::fSplB
private

Definition at line 64 of file ROCCalc.h.

◆ fSpleffBvsS

TSpline* TMVA::ROCCalc::fSpleffBvsS
private

Definition at line 66 of file ROCCalc.h.

◆ fSplmvaCumB

TSpline * TMVA::ROCCalc::fSplmvaCumB
private

spline of cumulated mva distributions

Definition at line 65 of file ROCCalc.h.

◆ fSplmvaCumS

TSpline* TMVA::ROCCalc::fSplmvaCumS
private

Definition at line 65 of file ROCCalc.h.

◆ fSplS

TSpline* TMVA::ROCCalc::fSplS
private

Definition at line 64 of file ROCCalc.h.

◆ fStatus

Bool_t TMVA::ROCCalc::fStatus
private

false if is found some error in mvaS or mvaB

Definition at line 54 of file ROCCalc.h.

◆ fUseSplines

Bool_t TMVA::ROCCalc::fUseSplines
private

Definition at line 57 of file ROCCalc.h.

◆ fXmax

Float_t TMVA::ROCCalc::fXmax
private

min and max of the mva distribution

Definition at line 61 of file ROCCalc.h.

◆ fXmin

Float_t TMVA::ROCCalc::fXmin
private

Definition at line 61 of file ROCCalc.h.

◆ inveffBvsS

TH1D* TMVA::ROCCalc::inveffBvsS
private

Definition at line 73 of file ROCCalc.h.

◆ rejBvsS

TH1D* TMVA::ROCCalc::rejBvsS
private

Definition at line 72 of file ROCCalc.h.

Libraries for TMVA::ROCCalc:
[legend]
The documentation for this class was generated from the following files: