Class to handle the fitting of toy dataset for limit production. More...

#include <ToyFitterExclusion.h>

Inheritance diagram for ToyFitterExclusion:

Public Member Functions
	ToyFitterExclusion (TString CollectionName)
	constructor, CollectionName is the name of the tree collection to fit, by convention set by ToyGenerator this is also the name of the file.

double	computeTS (double mu)
	returns the Test Statistic q_tilde(mu) for a value of mu_test. More...

void	fit (double mu, int stopAt=-999)
	given as input a file with many data toys, it writes a tree with fit outputs More...

void	setTheLikelihood (ProfileLikelihood *like)
	set the likelihood to fit

TGraphAsymmErrors	computeTSDistros (TString fileName, double *mu_list, int mu_size)
	produces a TGraph with the 90% quantiles of the alternate Hypothesis. More...

TGraphAsymmErrors	computeTSDistros (TTree tree, double mu_list, int mu_size)
	produces a TGraph with the 90% quantiles of the alternate Hypothesis. More...

void	spitTheLimit (TGraphAsymmErrors *ninety_quantiles, int sopAt=-999)

void	setInputDir (TString path)
	set path to input file (only dir no file name)

void	setOutputDir (TString path)
	set the path to dir in which the fits are stored otherwise is current

void	setRandomizeMeasure (bool doOrNot)

void	setTreeName (TString newName)

void	setSeed (ULong_t seed)

void	setCalibrationTreeName (TString newName)

void	setOutputSuffix (TString name)

void	setGeneration (int generation)
	set the Generation, this info will be available in the generated tree (so that u can hadd them), it is optional and non ncecessary.

Public Member Functions inherited from errorHandler
	errorHandler (TString name)

void	Error (TString functionName, TString message)
	Send an Error message to the user. More...

void	Warning (TString functionName, TString message)
	Send a Warning message to the user. More...

void	Info (TString functionName, TString message)
	Send an Info message to the user. More...

void	Debug (TString functionName, TString message)
	Send a Debug message to the user. More...

void	setPrintLevel (int level)
	Set the local printing level (for this instance only) More...

int	getPrintLevel ()
	Return the current print level for this instance. remind that if you set with setPrintLevel this will override the global print level.

Additional Inherited Members
Public Attributes inherited from errorHandler
int	localPrintLevel

TString	className

Static Public Attributes inherited from errorHandler
static int	globalPrintLevel = 1

Detailed Description

Class to handle the fitting of toy dataset for limit production.

Produces a Tree with all the post fit values.
Produces distributions of f(mu=mu_true | H_mu) and the 90% quantiles for a set of mu values.
Computes sensitivity + bands
Given a science dataset it compute the limit.

Member Function Documentation

◆ computeTS()

double ToyFitterExclusion::computeTS ( double mu )

returns the Test Statistic q_tilde(mu) for a value of mu_test.

It is a reproduction of q_tilde test statistic from equation 16 of: https://arxiv.org/abs/1007.1727 this is suitable for upper limit in which the parameter of interest must be >0. For it to work correctly is important to set up the likelihood beforehand.

Parameters

mu	the signal strenght for conditional fit.

◆ computeTSDistros() [1/2]

TGraphAsymmErrors ToyFitterExclusion::computeTSDistros	(	TString	fileName,
		double *	mu_list,
		int	mu_size
	)

produces a TGraph with the 90% quantiles of the alternate Hypothesis.

it is used to compute limits. The TH1F of the test statistic f(q_mu | H_mu) are stored in a root file toghether with the TGraph of 90% quantiles.

Parameters

fileName	takes as input the hadded post fit output from fit() of this class.
mu_list	is the list of true mu hypothesis present in the file tree.
mu_size	is the size of the previous list.

◆ computeTSDistros() [2/2]

TGraphAsymmErrors ToyFitterExclusion::computeTSDistros	(	TTree *	tree,
		double *	mu_list,
		int	mu_size
	)

produces a TGraph with the 90% quantiles of the alternate Hypothesis.

it is used to compute limits. The TH1F of the test statistic f(q_mu | H_mu) are stored in a root file toghether with the TGraph of 90% quantiles.

Parameters

tree	takes as input the tree of post fit output from fit() of this class.
mu_list	is the list of true mu hypothesis present in the file tree.
mu_size	is the size of the previous list.

◆ fit()

void ToyFitterExclusion::fit	(	double	mu,
		int	stopAt = `-999`
	)

given as input a file with many data toys, it writes a tree with fit outputs

Note that it expects a toy file generated with naming convention of ToyGenerator. The output tree contains: conditional and unconditional nuisance parameter post fit values, and the test statistic. This is supposed to be used for the alternative hypotesis fits, to be able to generate aftewards the graph of 90th percentiles. mu: the signal strenght of the conditional fit. soptAt: optional, the number of toy in file you want to fit. NOTE: by deafulte this will randomize the NP measures during fit for each toy, setRandomizeMeasure(false) if you don't want.

◆ spitTheLimit()

void ToyFitterExclusion::spitTheLimit	(	TGraphAsymmErrors *	ninety_quantiles,
		int	sopAt = `-999`
	)

given a input file with N null Hypo toy trees it produce an out tree containing post fits and limit.

note that this can be used also with just a signle tree, for real data limit production.

Parameters

ninety_quantiles	graph of 90% quantiles as produced from computeTSDistros().
TreeName	prefix name of the tree list you fit. To se the input file file use setPathToFile() method. soptAt: optional, the number of toy in file you want to fit. NOTE: by deafulte this will randomize the NP measures during fit for each toy, setRandomizeMeasure(false) if you don't want.

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

SOURCE/src/ToyFitterExclusion.h
SOURCE/src/ToyFitterExclusion.cxx

Public Member Functions

Additional Inherited Members

Detailed Description

Member Function Documentation

◆ computeTS()

◆ computeTSDistros() [1/2]

◆ computeTSDistros() [2/2]

◆ fit()

◆ spitTheLimit()