RooFit::TestStatistics Namespace Reference

Namespace for new RooFit test statistic calculation. More...

Classes
class	ConstantTermsOptimizer
	Analyzes a function given a dataset/observables for constant terms and caches those in the dataset. More...
class	LikelihoodGradientJob
class	LikelihoodGradientWrapper
	Virtual base class for implementation of likelihood gradient calculation strategies. More...
class	LikelihoodJob
class	LikelihoodSerial
	Serial likelihood calculation strategy implementation. More...
class	LikelihoodWrapper
	Virtual base class for implementation of likelihood calculation strategies. More...
class	MinuitFcnGrad
	Minuit-RooMinimizer interface which synchronizes parameter data and coordinates evaluation of likelihood (gradient) values. More...
class	NLLFactory
class	RooAbsL
class	RooBinnedL
class	RooRealL
	RooAbsReal that wraps RooAbsL likelihoods for use in RooFit outside of the RooMinimizer context. More...
class	RooSubsidiaryL
class	RooSumL
	Likelihood class that sums over multiple -log components. More...
class	RooUnbinnedL
struct	WrapperCalculationCleanFlags
	For communication with wrappers, an instance of this struct must be shared between them and MinuitFcnGrad. More...

Enumerations
enum class	LikelihoodGradientMode { multiprocess }
enum class	LikelihoodMode { serial , multiprocess }
enum class	LikelihoodType { unbinned , binned , subsidiary , sum }
enum class	OffsettingMode { legacy , full }
	Previously, offsetting was only implemented for RooNLLVar components of a likelihood, not for RooConstraintSum terms. More...

Functions
std::unique_ptr< RooAbsL >	buildLikelihood (RooAbsPdf *pdf, RooAbsData *data)
	Delegating function to build a likelihood without additional arguments.
std::ostream &	operator<< (std::ostream &out, const LikelihoodJob::update_state_mode value)

Detailed Description

Namespace for new RooFit test statistic calculation.

RooFit::TestStatistics contains a major refactoring of the RooAbsTestStatistic-RooAbsOptTestStatistic-RooNLLVar inheritance tree into:

1. statistics-based classes on the one hand;
2. calculation/evaluation/optimization based classes on the other hand.

The likelihood is the central unit on the statistics side. The RooAbsL class is implemented for four kinds of likelihoods: binned, unbinned, "subsidiary" (an optimization for numerical stability that gathers components like global observables) and "sum" (over multiple components of the other types). These classes provide ways to compute their components in parallelizable chunks that can be used by the calculator classes as they see fit.

On top of the likelihood classes, we also provide for convenience a likelihood builder buildLikelihood, as a free function in the namespace. This function analyzes the pdf and automatically constructs the proper likelihood, built up from the available RooAbsL subclasses.

The calculator "Wrapper" classes are abstract interfaces. These can be implemented for different kinds of algorithms, or with different kinds of optimization "back-ends" in mind. In an upcoming PR, we will introduce the fork-based multi-processing implementation based on RooFit::MultiProcess. Other possible implementations could use the GPU or external tools like TensorFlow.

The coupling of all these classes to RooMinimizer is made via the MinuitFcnGrad class, which owns the Wrappers that calculate the likelihood components.

Enumeration Type Documentation

LikelihoodGradientMode

enum class RooFit::TestStatistics::LikelihoodGradientMode

strong

Enumerator
multiprocess

Definition at line 33 of file LikelihoodGradientWrapper.h.

LikelihoodMode

enum class RooFit::TestStatistics::LikelihoodMode

strong

Enumerator
serial
multiprocess

Definition at line 50 of file LikelihoodWrapper.h.

LikelihoodType

enum class RooFit::TestStatistics::LikelihoodType

strong

Enumerator
unbinned
binned
subsidiary
sum

Definition at line 48 of file LikelihoodWrapper.h.

OffsettingMode

enum class RooFit::TestStatistics::OffsettingMode

strong

Previously, offsetting was only implemented for RooNLLVar components of a likelihood, not for RooConstraintSum terms.

To emulate this behavior, use OffsettingMode::legacy. To also offset the RooSubsidiaryL component (equivalent of RooConstraintSum) of RooSumL likelihoods, use OffsettingMode::full.

Enumerator
legacy
full

Definition at line 56 of file LikelihoodWrapper.h.

Function Documentation

buildLikelihood()

std::unique_ptr< RooAbsL > RooFit::TestStatistics::buildLikelihood ( RooAbsPdf *  pdf, RooAbsData *  data  )

inline

Delegating function to build a likelihood without additional arguments.

Definition at line 55 of file buildLikelihood.h.

operator<<()

std::ostream & RooFit::TestStatistics::operator<<	(	std::ostream &	out,
		const LikelihoodJob::update_state_mode	value
	)

Definition at line 289 of file LikelihoodJob.cxx.