MCEngine.h File Reference

Monte Carlo propagation of nuisance-parameter uncertainties. More...

#include <vector>
#include <random>
#include <string>
#include <memory>
#include "INuisanceSampler.h"
#include "ports/IModel.h"
#include "Statistics.h"
#include "GaussianApprox.h"
#include "Indexing.h"
#include "StatisticProgress.h"

Include dependency graph for MCEngine.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
struct	MCConfig
	Runtime configuration for Monte Carlo nuisance propagation. More...
struct	MCRealization
	Raw Monte Carlo samples accepted by the engine. More...
struct	MCObservableCovariance
	Empirical observable covariance and its inverse. More...
struct	MCResult
	Complete output of a Monte Carlo propagation run. More...
class	MonteCarloEngine
	Samples nuisance parameters and propagates them through a model. More...

Functions
MCObservableCovariance	covariance_from_obs_samples (const ObsSamples &S, const std::vector< BinnedObservableId > &ids, double ridge_rel=1e-8, double ridge_abs=1e-12)
	Builds a regularized empirical covariance matrix from observable samples.
std::vector< BinnedObservableId >	covariance_ids_from_first_sample (const ObsSamples &S)
	Extracts the observable ordering from the first Monte Carlo sample.

Detailed Description

Monte Carlo propagation of nuisance-parameter uncertainties.

The Monte Carlo engine samples nuisance parameters, evaluates model predictions, summarizes the resulting observable distributions, and builds a regularized observable covariance matrix.

See also

INuisanceSampler

IModel

GaussianSummary

Definition in file MCEngine.h.

Function Documentation

covariance_from_obs_samples()

MCObservableCovariance covariance_from_obs_samples	(	const ObsSamples &	S,
		const std::vector< BinnedObservableId > &	ids,
		double	ridge_rel = 1e-8,
		double	ridge_abs = 1e-12
	)

Builds a regularized empirical covariance matrix from observable samples.

The samples are read in the order specified by ids. The empirical covariance is converted to a dimensionless correlation matrix, a ridge of max(ridge_rel, ridge_abs) is added to its diagonal, and the inverse is transformed back to covariance units. The returned covariance itself remains the raw empirical estimate.

Parameters

S	Monte Carlo observable samples.
ids	Observable identifiers defining the covariance ordering.
ridge_rel	Relative ridge factor.
ridge_abs	Minimum dimensionless ridge floor.

Returns

Empirical covariance, inverse covariance and column means.

Exceptions

std::invalid_argument

if the sample set is empty, if ids is empty, or if fewer than two samples are provided.

Definition at line 102 of file MCEngine.cpp.

covariance_ids_from_first_sample()

std::vector< BinnedObservableId > covariance_ids_from_first_sample

(

const ObsSamples &

S

)

Extracts the observable ordering from the first Monte Carlo sample.

Parameters

S	Monte Carlo observable samples.

Returns

Observable identifiers present in the first sample.

Exceptions

std::invalid_argument

if S is empty.

Definition at line 164 of file MCEngine.cpp.