A class for fitting histograms with functions. More...

#include <BCHistogramFitter.h>

Inheritance diagram for BCHistogramFitter:

Collaboration diagram for BCHistogramFitter:

Public Member Functions
Constructors and destructors
BCHistogramFitter (const TH1 &hist, const TF1 &func, const std::string &name="histogram_fitter_model")
Constructor. More...

virtual ~BCHistogramFitter ()
The default destructor. More...

Member functions (get)
const TH1 & GetHistogram ()

Member functions (miscellaneous methods)
virtual double LogLikelihood (const std::vector< double > &parameters)
The log of the conditional probability. More...

virtual void Fit ()
Performs the fit. More...

void DrawFit (const std::string &options="HIST", bool flaglegend=false)
Draw the fit in the current pad. More...

double CalculatePValueFast (const std::vector< double > &par, unsigned nIterations=100000)
Calculate the p-value using fast-MCMC and the likelihood as test statistic. More...

double CalculatePValueLikelihood (const std::vector< double > &par)
Calculate the p-value using approximate chi^2 distribution of scaled likelihood. More...

double CalculatePValueLeastSquares (const std::vector< double > &par, bool weightExpect=true)
Calculate the p-value using approximate chi^2 distribution of squared difference for conventional weights. More...

Public Member Functions inherited from BCFitter
virtual void MCMCUserInitialize ()
Create enough TF1 copies for thread safety.

BCFitter (const TF1 &f, const std::string &name="fitter_model")
Constructor. More...

virtual ~BCFitter ()=0
The default destructor. More...

TF1 & GetFitFunction ()
Get fit function. More...

TH2 * GetGraphicalErrorBandXY (double level=.68, int nsmooth=0, bool overcoverage=true) const

const TH2 & GetErrorBandXY () const

std::vector< double > GetErrorBand (double level) const
Returns a vector of y-values at a certain probability level. More...

TGraph * GetErrorBandGraph (double level1, double level2) const

TGraph * GetFitFunctionGraph (const std::vector< double > &parameters)

TGraph * GetFitFunctionGraph ()

TGraph * GetFitFunctionGraph (const std::vector< double > &parameters, double xmin, double xmax, int n=1000)

void FixDataAxis (unsigned int index, bool fixed)
Toggle the data axis defined by `index` to be fixed. More...

bool GetFixedDataAxis (unsigned int index) const

double GetPValue () const

void SetErrorBandContinuous (bool flag)
Sets the error band flag to continuous function.

void SetErrorBandExtensionLowEdgeX (double extension)
Extends the lower x Edge of th errorband by -extension.

void SetErrorBandExtensionUpEdgeX (double extension)
Extends the lower x Edge of th errorband by +extension.

void SetErrorBandExtensionLowEdgeY (double extension)
Extends the lower y Edge of th errorband by -extension.

void SetErrorBandExtensionUpEdgeY (double extension)
Extends the lower y Edge of th errorband by +extension.

void SetFillErrorBand (bool flag=true)
Turn on or off the filling of the error band during the MCMC run. More...

void UnsetFillErrorBand ()
Turn off filling of the error band during the MCMC run. More...

void SetFitFunctionIndexX (int index)
Sets index of the x values in function fits. More...

void SetFitFunctionIndexY (int index)
Sets index of the y values in function fits. More...

void SetFitFunctionIndices (int indexx, int indexy)
Sets indices of the x and y values in function fits. More...

void SetFlagIntegration (bool flag)
Sets the flag for integration. More...

virtual double FitFunction (const std::vector< double > &x, const std::vector< double > &parameters)
Defines a fit function. More...

double Integral (const std::vector< double > &parameters, double xmin, double xmax)
Compute the integral of the fit function between xmin and xmax. More...

virtual void MCMCUserIterationInterface ()
Overloaded from BCEngineMCMC.

virtual void MarginalizePreprocess ()
Overloaded from BCIntegrate. More...

void FillErrorBand ()
Fill error band histogram for current iteration. More...

void PrintShortFitSummary ()
Prints a short summary of the fit results on the screen. More...

Public Member Functions inherited from BCModel
BCModel (const std::string &name="model")
Default constructor. More...

BCModel (const BCModel &bcmodel)
Copy constructor. More...

BCModel (const std::string &filename, const std::string &name, bool loadObservables=true)
Read in MCMC constructor. More...

BCModel & operator= (const BCModel &)
Copy-assignment operator.

virtual ~BCModel ()
Destructor. More...

BCDataSet * GetDataSet ()

unsigned GetNDataPoints () const

int GetNDoF () const

virtual BCPriorModel * GetPriorModel (bool prepare=true, bool call_likelihood=false)

virtual BCH1D GetPrior (unsigned index)
Get prior of a variable as a BCH1D. More...

virtual BCH2D GetPrior (unsigned index1, unsigned index2)
Get prior of a pair of variables as a BCH2D. More...

BCH1D & GetBCH1DPriorDrawingOptions ()

BCH2D & GetBCH2DPriorDrawingOptions ()

BCH1D & GetBCH1DPosteriorDrawingOptions ()

BCH2D & GetBCH2DPosteriorDrawingOptions ()

bool GetDrawPriorFirst () const

void SetDataSet (BCDataSet *dataset)
Sets the data set. More...

void SetKnowledgeUpdateDrawingStyle (BCAux::BCKnowledgeUpdateDrawingStyle style=BCAux::kKnowledgeUpdateDefaultStyle)
Set default drawing options for knowledge update plots. More...

void SetDrawPriorFirst (bool b=true)
Set drawing of prior first (true) or posterior first (false) for knowledge update plots. More...

virtual double APrioriProbability (const std::vector< double > &parameters)
Returns the prior probability. More...

virtual double LogAPrioriProbability (const std::vector< double > &parameters)
Returns natural logarithm of the prior probability. More...

virtual double Likelihood (const std::vector< double > &params)
Returns the likelihood. More...

virtual double ProbabilityNN (const std::vector< double > &params)
Returns the likelihood times prior probability given a set of parameter values. More...

virtual double LogProbabilityNN (const std::vector< double > &parameters)
Returns the natural logarithm of likelihood times prior probability given a set of parameter values. More...

virtual double Probability (const std::vector< double > &parameters)
Returns the a posteriori probability given a set of parameter values. More...

virtual double LogProbability (const std::vector< double > &parameters)
Returns natural logarithm of the a posteriori probability given a set of parameter values. More...

virtual double SamplingFunction (const std::vector< double > &parameters)
Sampling function used for importance sampling. More...

virtual double Eval (const std::vector< double > &parameters)
Overloaded function to evaluate integral. More...

virtual double LogEval (const std::vector< double > &parameters)
Overloaded function to evaluate integral. More...

virtual void InitializeMarkovChainTree (bool replacetree=false, bool replacefile=false)
Initialize the trees containing the Markov chains and parameter info. More...

double HessianMatrixElement (unsigned index1, unsigned index2, const std::vector< double > &point)
Calculates the matrix element of the Hessian matrix. More...

void PrintShortFitSummary ()
Prints a short summary of the fit results on the screen. More...

void PrintHessianMatrix (std::vector< double > parameters)
Prints matrix elements of the Hessian matrix. More...

virtual unsigned PrintKnowledgeUpdatePlots (const std::string &filename, unsigned hdiv=1, unsigned vdiv=1, bool call_likelihood=false)
Print a comparison of the prior knowledge to the posterior knowledge for each parameter. More...

Public Member Functions inherited from BCIntegrate
BCIntegrate (const std::string &name="model")
Default constructor.

BCIntegrate (const std::string &filename, const std::string &name, bool loadObservables=true)
Read in MCMC constructor. More...

BCIntegrate (const BCIntegrate &other)
Copy constructor.

BCIntegrate & operator= (const BCIntegrate &)
Copy-assignment operator.

virtual ~BCIntegrate ()
Destructor.

double GetIntegral () const

BCIntegrate::BCOptimizationMethod GetOptimizationMethod () const

BCIntegrate::BCIntegrationMethod GetIntegrationMethod () const

BCIntegrate::BCMarginalizationMethod GetMarginalizationMethod () const

BCIntegrate::BCSASchedule GetSASchedule () const

void GetRandomVectorInParameterSpace (std::vector< double > &x) const
Fills a vector of random numbers x[i] between fMin[i] and fMax[i] into a vector. More...

double GetRandomPoint (std::vector< double > &x)
Fills a vector of (flat) random numbers in the limits of the parameters and returns the probability at that point. More...

int GetNIterationsMin () const

int GetNIterationsMax () const

int GetNIterationsPrecisionCheck () const

int GetNIterations () const

double GetRelativePrecision () const

double GetAbsolutePrecision () const

BCCubaMethod GetCubaIntegrationMethod () const

const BCCubaOptions::Vegas & GetCubaVegasOptions () const

const BCCubaOptions::Suave & GetCubaSuaveOptions () const

const BCCubaOptions::Divonne & GetCubaDivonneOptions () const

const BCCubaOptions::Cuhre & GetCubaCuhreOptions () const

TH1 * GetSlice (std::vector< unsigned > indices, unsigned &nIterations, double &log_max_val, const std::vector< double > parameters=std::vector< double >(0), int nbins=0, bool normalize=true)
Returns a one-dimensional slice of the pdf at the point and along a specific direction. More...

TH1 * GetSlice (const std::string &name, unsigned &nIterations, double &log_max_val, const std::vector< double > parameters=std::vector< double >(0), int nbins=0, bool normalize=true)
Returns a one-dimensional slice of the pdf at the point and along a specific direction. More...

TH1 * GetSlice (unsigned index, unsigned &nIterations, double &log_max_val, const std::vector< double > parameters=std::vector< double >(0), int nbins=0, bool normalize=true)
Returns a one-dimensional slice of the pdf at the point and along a specific direction. More...

TH2 * GetSlice (const std::string &name1, const std::string &name2, unsigned &nIterations, double &log_max_val, const std::vector< double > parameters=std::vector< double >(0), int nbins=0, bool normalize=true)
Returns a two-dimensional slice of the pdf at the point and along two specified directions. More...

TH2 * GetSlice (unsigned index1, unsigned index2, unsigned &nIterations, double &log_max_val, const std::vector< double > parameters=std::vector< double >(0), int nbins=0, bool normalize=true)
Returns a two-dimensional slice of the pdf at the point and along two specified directions. More...

double GetError () const

TMinuitMinimizer & GetMinuit ()

double GetSAT0 () const
Returns the Simulated Annealing starting temperature. More...

double GetSATmin () const
Returns the Simulated Annealing threshhold temperature. More...

virtual const std::vector< double > & GetBestFitParameters () const

const std::vector< double > & GetBestFitParameterErrors () const
Returns the set of errors on the values of the parameters at the mode.

double GetLogMaximum () const
Returns the posterior at the mode. More...

void SetFlagIgnorePrevOptimization (bool flag)

void SetOptimizationMethod (BCIntegrate::BCOptimizationMethod method)

void SetIntegrationMethod (BCIntegrate::BCIntegrationMethod method)

void SetMarginalizationMethod (BCIntegrate::BCMarginalizationMethod method)

void SetSASchedule (BCIntegrate::BCSASchedule schedule)

void SetNIterationsMin (int niterations)

void SetNIterationsMax (int niterations)

void SetNIterationsPrecisionCheck (int niterations)

void SetRelativePrecision (double relprecision)

void SetAbsolutePrecision (double absprecision)
Set absolute precision of the numerical integation.

void SetCubaIntegrationMethod (BCCubaMethod type)
Set Cuba integration method.

void SetCubaOptions (const BCCubaOptions::Vegas &options)
Set options for CUBA's Vegas. More...

void SetCubaOptions (const BCCubaOptions::Suave &options)
Set options for CUBA's Suave. More...

void SetCubaOptions (const BCCubaOptions::Divonne &options)
Set options for CUBA's Divonne. More...

void SetCubaOptions (const BCCubaOptions::Cuhre &options)
Set options for CUBA's Cuhre. More...

void SetSAT0 (double T0)
Set starting temperature for Simulated Annealing. More...

void SetSATmin (double Tmin)
Set threshold temperature for Simulated Annealing. More...

double Normalize ()
Performs integration. More...

double Integrate (BCIntegrationMethod intmethod)
Does the integration over the un-normalized probability. More...

double Integrate ()
Perform the integration. More...

double Integrate (BCIntegrationMethod type, tRandomizer randomizer, tEvaluator evaluator, tIntegralUpdater updater, std::vector< double > &sums)
Does the integration over the un-normalized probability. More...

double EvaluatorMC (std::vector< double > &sums, const std::vector< double > &point, bool &accepted)
Evaluates integrator.

int MarginalizeAll ()
Marginalize all probabilities wrt. More...

int MarginalizeAll (BCMarginalizationMethod margmethod)
Marginalize all probabilities wrt. More...

virtual void MarginalizePostprocess ()
Method executed after marginalization. More...

std::vector< double > FindMode (std::vector< double > start=std::vector< double >())
Do the mode finding using a method set via SetOptimizationMethod. More...

std::vector< double > FindMode (BCIntegrate::BCOptimizationMethod optmethod, std::vector< double > start=std::vector< double >())
Find mode using a specific method. More...

double SATemperature (double t) const
Temperature annealing schedule for use with Simulated Annealing. More...

double SATemperatureBoltzmann (double t) const
Temperature annealing schedule for use with Simulated Annealing. More...

double SATemperatureCauchy (double t) const
Temperature annealing schedule for use with Simulated Annealing. More...

virtual double SATemperatureCustom (double t) const
Temperature annealing schedule for use with Simulated Annealing. More...

std::vector< double > GetProposalPointSA (const std::vector< double > &x, int t) const
Generates a new state in a neighbourhood around x that is to be accepted or rejected by the Simulated Annealing algorithm. More...

std::vector< double > GetProposalPointSABoltzmann (const std::vector< double > &x, int t) const
Generates a new state in a neighbourhood around x that is to be accepted or rejected by the Simulated Annealing algorithm. More...

std::vector< double > GetProposalPointSACauchy (const std::vector< double > &x, int t) const
Generates a new state in a neighbourhood around x that is to be accepted or rejected by the Simulated Annealing algorithm. More...

virtual std::vector< double > GetProposalPointSACustom (const std::vector< double > &x, int t) const
Generates a new state in a neighbourhood around x that is to be accepted or rejected by the Simulated Annealing algorithm. More...

std::vector< double > SAHelperGetRandomPointOnHypersphere () const
Generates a uniform distributed random point on the surface of a fNvar-dimensional Hypersphere. More...

double SAHelperGetRadialCauchy () const
Generates the radial part of a n-dimensional Cauchy distribution. More...

double SAHelperSinusToNIntegral (int dim, double theta) const
Returns the Integral of sin^dim from 0 to theta. More...

virtual void ResetResults ()
Reset all information on the best-fit parameters. More...

std::string DumpIntegrationMethod (BCIntegrationMethod type) const
Return string with the name for a given integration type. More...

std::string DumpCurrentIntegrationMethod () const
Return string with the name for the currently set integration type. More...

std::string DumpUsedIntegrationMethod () const
Return string with the name for the previously used integration type. More...

std::string DumpMarginalizationMethod (BCMarginalizationMethod type) const
Return string with the name for a given marginalization type. More...

std::string DumpCurrentMarginalizationMethod () const
Return string with the name for the currently set marginalization type. More...

std::string DumpUsedMarginalizationMethod () const
Return string with the name for the marginalization type used. More...

std::string DumpOptimizationMethod (BCOptimizationMethod type) const
Return string with the name for a given optimization type. More...

std::string DumpCurrentOptimizationMethod () const
Return string with the name for the currently set optimization type. More...

std::string DumpUsedOptimizationMethod () const
Return string with the name for the optimization type used to find the current mode. More...

std::string DumpCubaIntegrationMethod (BCCubaMethod type) const
Return string with the name for a given Cuba integration type. More...

std::string DumpCubaIntegrationMethod () const
Return string with the name for the currently set Cuba integration type. More...

void SetBestFitParameters (const std::vector< double > &x)
Set best fit parameters values. More...

void SetBestFitParameters (const std::vector< double > &x, const double &new_value, double &old_value)

bool CheckMarginalizationAvailability (BCMarginalizationMethod type)
Check availability of integration routine for marginalization. More...

bool CheckMarginalizationIndices (TH1 *hist, const std::vector< unsigned > &index)
Check that indices of parameters to marginalize w/r/t are correct.

double IntegrateLaplace ()
Integrate using the Laplace approximation. More...

Public Member Functions inherited from BCEngineMCMC
BCEngineMCMC (const std::string &name="model")
Default constructor. More...

BCEngineMCMC (const BCEngineMCMC &enginemcmc)
Copy constructor. More...

BCEngineMCMC (const std::string &filename, const std::string &name, bool loadObservables=true)
Read in MCMC constructor. More...

BCEngineMCMC & operator= (const BCEngineMCMC &)
Copy-assignment operator.

virtual ~BCEngineMCMC ()
Destructor. More...

const std::string & GetName () const

const std::string & GetSafeName () const

unsigned GetNChains () const

unsigned GetNLag () const

int GetCurrentIteration () const

unsigned GetCurrentChain () const

int GetNIterationsConvergenceGlobal () const

unsigned GetNIterationsPreRun () const

unsigned GetNIterationsPreRunMin () const

unsigned GetNIterationsPreRunMax () const

unsigned GetNIterationsRun () const

unsigned GetNIterationsPreRunCheck () const

unsigned GetPreRunCheckClear ()

double GetMinimumEfficiency () const

double GetMaximumEfficiency () const

double GetScaleFactorLowerLimit () const

double GetScaleFactorUpperLimit () const

const std::vector< std::vector< double > > & GetScaleFactors () const

const ChainState & GetChainState (unsigned c) const

const std::vector< double > & Getx (unsigned c) const

double Getx (unsigned c, unsigned p) const

double GetLogProbx (unsigned c) const

BCEngineMCMC::Phase GetPhase () const

BCEngineMCMC::InitialPositionScheme GetInitialPositionScheme () const

unsigned GetInitialPositionAttemptLimit () const

bool GetProposeMultivariate () const

double GetProposalFunctionDof () const

unsigned GetMultivariateCovarianceUpdates () const

double GetMultivariateCovarianceUpdateLambda () const

double GetMultivariateEpsilon () const

double GetMultivariateScaleMultiplier () const

double GetRValueParametersCriterion () const

const std::vector< double > & GetRValueParameters () const

double GetRValueParameters (unsigned index) const

bool GetCorrectRValueForSamplingVariability () const
Flag for correcting convergence checking for initial sampling variability. More...

bool GetFlagRun () const

TTree * GetMarkovChainTree () const
Retrieve the tree containing the Markov chain. More...

TTree * GetParameterTree () const
Retrieve the tree containing the parameter information. More...

TFile * GetOutputFile () const
Retrieve output file for MCMC. More...

const BCEngineMCMC::Statistics & GetStatistics () const
Get combined statistics for all chains. More...

const BCEngineMCMC::Statistics & GetStatistics (unsigned c) const
Get MCMC statistics for one chain. More...

const std::vector< BCEngineMCMC::Statistics > & GetStatisticsVector () const
Get vector of MCMC statistics for each chain separately. More...

bool GetRescaleHistogramRangesAfterPreRun () const

double GetHistogramRescalePadding () const

virtual std::vector< unsigned > GetH1DPrintOrder () const

virtual std::vector< std::pair< unsigned, unsigned > > GetH2DPrintOrder () const

bool MarginalizedHistogramExists (unsigned index) const

bool MarginalizedHistogramExists (unsigned index1, unsigned index2) const

TH1 * GetMarginalizedHistogram (const std::string &name) const
Obtain the individual marginalized distributions with respect to one parameter as a ROOT TH1. More...

TH1 * GetMarginalizedHistogram (unsigned index) const
Obtain the individual marginalized distributions with respect to one parameter as a ROOT TH1. More...

TH2 * GetMarginalizedHistogram (const std::string &name1, const std::string &name2) const
Obtain the individual marginalized distributions with respect to two parameters as a ROOT TH2. More...

TH2 * GetMarginalizedHistogram (unsigned index1, unsigned index2) const
Obtain the individual marginalized distributions with respect to two parameters as a ROOT TH2. More...

BCH1D GetMarginalized (const std::string &name) const
Obtain the individual marginalized distributions with respect to one parameter. More...

BCH1D GetMarginalized (unsigned index) const
Obtain the individual marginalized distributions with respect to one parameter. More...

BCH2D GetMarginalized (const std::string &name1, const std::string &name2) const
Obtain the individual marginalized distributions with respect to two parameters. More...

BCH2D GetMarginalized (unsigned index1, unsigned index2) const
Obtain the individual marginalized distributions with respect to two parameters. More...

unsigned GetMaximumParameterNameLength (bool observables=true) const

BCVariable & GetVariable (unsigned index)

const BCVariable & GetVariable (unsigned index) const

unsigned GetNVariables () const

BCParameterSet & GetParameters ()

const BCParameterSet & GetParameters () const

BCParameter & GetParameter (unsigned index)

const BCParameter & GetParameter (unsigned index) const

BCParameter & GetParameter (const std::string &name)

const BCParameter & GetParameter (const std::string &name) const

unsigned GetNParameters () const

unsigned GetNFixedParameters () const

unsigned GetNFreeParameters () const

BCObservableSet & GetObservables ()

const BCObservableSet & GetObservables () const

BCObservable & GetObservable (unsigned index)

const BCObservable & GetObservable (unsigned index) const

BCObservable & GetObservable (const std::string &name)

const BCObservable & GetObservable (const std::string &name) const

unsigned GetNObservables () const

const std::vector< double > & GetLocalModes (bool force_recalculation=false)

bool GetReuseObservables () const

BCH1D & GetBCH1DdrawingOptions ()

BCH2D & GetBCH2DdrawingOptions ()

void SetName (const std::string &name)
Sets the name of the engine. More...

void SetScaleFactorLowerLimit (double l)
Set scale factor lower limit.

void SetScaleFactorUpperLimit (double l)
Set scale factor upper limit.

void SetInitialScaleFactors (const std::vector< double > &scale)
Set the initial scale factors for the factorized proposal function. More...

void SetNChains (unsigned n)
Sets the number of Markov chains which are run in parallel. More...

void SetNLag (unsigned n)
Sets the lag of the Markov chains.

void SetNIterationsPreRunMax (unsigned n)
Sets the maximum number of iterations in the pre-run. More...

void SetNIterationsRun (unsigned n)
Sets the number of iterations. More...

void SetNIterationsPreRunMin (unsigned n)
Sets the minimum number of iterations in the pre-run.

void SetNIterationsPreRunCheck (unsigned n)
Sets the number of iterations between scale adjustments and convergence checks in the pre-run. More...

void SetPreRunCheckClear (unsigned n)
Sets the number of prerun checks to make inbetween statistics clearing. More...

void SetMinimumEfficiency (double efficiency)
Sets the minimum efficiency required for a chain. More...

void SetMaximumEfficiency (double efficiency)
Sets the maximum efficiency required for a chain. More...

void SetRandomSeed (unsigned seed)
Set the random number seed.

void SetInitialPositions (const std::vector< double > &x0s)
Sets the initial positions for all chains. More...

void SetInitialPositions (const std::vector< std::vector< double > > &x0s)
Sets the initial positions for all chains. More...

void SetInitialPositionScheme (BCEngineMCMC::InitialPositionScheme scheme)
Sets flag which defines initial position. More...

void SetInitialPositionAttemptLimit (unsigned n)
Sets maximum number of attempts to find a valid initial position. More...

void SetProposeMultivariate (bool flag)
Set `flag` to `true` to turn on the multivariate proposal for MCMC based on (Haario et al., 2001) where the covariance is learned from the prerun. More...

void SetProposalFunctionDof (double dof=1)
Set the degree of freedom of the proposal function for MCMC. More...

void SetMultivariateCovarianceUpdateLambda (double l)
Set weighting for multivariate proposal function covariance update. More...

void SetMultivariateEpsilon (double epsilon)
Sets multivariate-proposal-function cholesky-decomposition nudge. More...

void SetMultivariateScaleMultiplier (double s)
Sets multivariate-proposal-function scale multiplier. More...

void SetFlagFillHistograms (bool flag)
Sets whether to fill histograms. More...

void SetFlagFillHistograms (bool flag_1d, bool flag_2d)
Sets the whether to fill histograms. More...

void SetFillHistogramParPar (unsigned x, unsigned y, bool flag=true)
Sets whether to fill particular H2 histogram: par(y) vs. More...

void SetFillHistogramParPar (const std::string &x, const std::string &y, bool flag=true)
Sets whether to fill particular H2 histogram: par(y) vs. More...

void SetFillHistogramParObs (unsigned x, unsigned y, bool flag=true)
Sets whether to fill particular H2 histogram: obs(y) vs. More...

void SetFillHistogramParObs (const std::string &x, const std::string &y, bool flag=true)
Sets whether to fill particular H2 histogram: obs(y) vs. More...

void SetFillHistogramObsObs (unsigned x, unsigned y, bool flag=true)
Sets whether to fill particular H2 histogram: obs(y) vs. More...

void SetFillHistogramObsObs (const std::string &x, const std::string &y, bool flag=true)
Sets whether to fill particular H2 histogram: obs(y) vs. More...

void SetFillHistogramObsPar (unsigned x, unsigned y, bool flag=true)
Sets whether to fill particular H2 histogram: par(y) vs. More...

void SetFillHistogramObsPar (const std::string &x, const std::string &y, bool flag=true)
Sets whether to fill particular H2 histogram: par(y) vs. More...

void SetFlagPreRun (bool flag)
Set if a (new) prerun should be performed. More...

void SetRValueParametersCriterion (double r)
Sets the parameter R-value criterion for convergence of all chains.

void SetCorrectRValueForSamplingVariability (bool flag=true)
Set flag to correct convergence checking for initial sampling variability. More...

void SetPrecision (BCEngineMCMC::Precision precision)
Set the precision for the MCMC run. More...

void SetPrecision (const BCEngineMCMC *other)
Copy precision for the MCMC run from other model. More...

void SetPrecision (const BCEngineMCMC &other)
Copy precision for the MCMC run from other model. More...

void SetNbins (unsigned int nbins)
Set the number of bins for the marginalized distribution of all parameters. More...

void SetReuseObservables (bool flag)

void SetRescaleHistogramRangesAfterPreRun (bool flag=true)
Set flag for rescaling histogram ranges after pre-run. More...

void SetHistogramRescalingPadding (double factor)
Set enlargement factor of range for when rescaling. More...

void WriteMarkovChain (bool flag)
Turn on/off writing of Markov chain to root file. More...

void WriteMarkovChainRun (bool flag)
Turn on/off writing of Markov chain to root file during run. More...

void WriteMarkovChainPreRun (bool flag)
Turn on/off writing of Markov chain to root file during prerun. More...

void WriteMarkovChain (const std::string &filename, const std::string &option, bool flag_run=true, bool flag_prerun=true)
Turn on writing of Markov chain to root file. More...

void SetPriorConstant (unsigned index)

void SetPriorConstant (const std::string &name)

void SetPrior (unsigned index, TF1 &f, bool logL=true)

void SetPrior (const std::string &name, TF1 &f, bool logL=true)

void SetPriorDelta (unsigned index, double value)

void SetPriorDelta (const std::string &name, double value)

void SetPriorGauss (unsigned index, double mean, double sigma)

void SetPriorGauss (const std::string &name, double mean, double sigma)

void SetPriorGauss (unsigned index, double mode, double sigma_below, double sigma_above)

void SetPriorGauss (const std::string &name, double mode, double sigma_below, double sigma_above)

void SetPrior (unsigned index, TH1 &h, bool interpolate=false)

void SetPrior (const std::string &name, TH1 &h, bool interpolate=false)

void SetPriorConstantAll ()

void WriteMarginalizedDistributions (const std::string &filename, const std::string &option, bool closeExistingFile=false)
Write marginalization histograms to file. More...

virtual void PrintSummary () const
Prints a summary to the logs. More...

void PrintParameters (const std::vector< double > &P, void(*output)(const std::string &)=BCLog::OutSummary) const
Print parameters. More...

unsigned PrintAllMarginalized (const std::string &filename, unsigned hdiv=1, unsigned vdiv=1) const
Print all marginalizations. More...

unsigned PrintParameterPlot (const std::string &filename, int npar=10, double interval_content=68e-2, std::vector< double > quantile_values=std::vector< double >(0), bool rescale_ranges=true) const
Print a summary plot for the parameters and user-defined observables. More...

bool DrawParameterPlot (unsigned i0, unsigned npar=0, double interval_content=68e-2, std::vector< double > quantile_values=std::vector< double >(0), bool rescale_ranges=true) const
Draw a summary plot for the parameters in the range provided to current pad. More...

bool PrintCorrelationMatrix (const std::string &filename="matrix.pdf") const
Print a correlation matrix for the parameters. More...

bool PrintCorrelationPlot (const std::string &filename="correlation.pdf", bool include_observables=true) const
Print a correlation plot for the parameters. More...

bool PrintParameterLatex (const std::string &filename) const
Print a LaTeX table of the parameters. More...

virtual void CreateHistograms (bool rescale_ranges=false)
Create histograms from parameter and observable sets. More...

virtual bool AddParameter (const std::string &name, double min, double max, const std::string &latexname="", const std::string &unitstring="")

virtual bool AddParameter (BCParameter &parameter)

virtual bool AddObservable (const std::string &name, double min, double max, const std::string &latexname="", const std::string &unitstring="")

virtual bool AddObservable (BCObservable &obs)

virtual void EvaluateObservables ()
Evaluates user-defined observables at current state of all chains and stores results in fMCMCState.

virtual void EvaluateObservables (unsigned chain)
Evaluates user-defined observables at current state of chain and stores results in fMCMCState. More...

virtual void CalculateObservables (const std::vector< double > &pars)
Evaluates user-defined observables. More...

virtual double ProposalFunction (unsigned ichain, unsigned ipar)
The default proposal function is a Breit-Wigner random walk. More...

bool GetProposalPointMetropolis (unsigned chain, std::vector< double > &x)
Return a proposal point for the Metropolis algorithm. More...

bool GetProposalPointMetropolis (unsigned chain, unsigned parameter, std::vector< double > &x)
Return a proposal point for the Metropolis algorithm. More...

bool GetNewPointMetropolis ()
Generate a new point using the Metropolis algorithm for all chains. More...

bool GetNewPointMetropolis (unsigned chain)
Generate a new point using the Metropolis algorithm for one chain. More...

bool GetNewPointMetropolis (unsigned chain, unsigned parameter)
Generate a new point using the Metropolis algorithm for one chain, varying only one parameter's value. More...

bool AcceptOrRejectPoint (unsigned chain, unsigned parameter)
Accept or rejects a point for a chain and updates efficiency. More...

void InChainFillHistograms (const ChainState &cs)
Fill marginalized distributions from a chain state.

void InChainFillHistograms ()
Fill marginalized distributions from all chain states.

void InChainFillTree (const ChainState &cs, unsigned chain_number)
Write a chain state to the tree.

void InChainFillTree ()
Write all chain states to the tree.

bool Metropolis ()
Runs Metropolis algorithm. More...

bool MetropolisPreRun ()
Runs a pre run for the Metropolis algorithm. More...

void MCMCInitialize ()
Resets all containers used in MCMC and initializes starting points. More...

virtual void MCMCCurrentPointInterface (const std::vector< double > &point, int ichain, bool accepted)
Interface allowing to execute arbitrary code for each new point of the MCMC whether it is accepted or not. More...

void LoadParametersFromTree (TTree *partree, bool loadObservables=true)
Load parameters and observables from tree. More...

void LoadMCMCParameters (TTree &partree)
Load MCMC parameters from parameter tree: nchains, proposal function type, scales. More...

virtual bool ParameterTreeMatchesModel (TTree *partree, bool checkObservables=true)
Check parameter tree against model. More...

void LoadMCMC (const std::string &filename, std::string mcmcTreeName="", std::string parameterTreeName="", bool loadObservables=true)
Load previous MCMC run. More...

void LoadMCMC (TTree mcmcTree, TTree parTree, bool loadObservables=true)
Load previous MCMC run. More...

bool ValidMCMCTree (TTree *tree, bool checkObservables=true) const
Check tree structure for MCMC tree. More...

bool ValidParameterTree (TTree *tree) const
Check tree structure for parameter tree. More...

void CloseOutputFile ()
Close the root output file. More...

virtual void Remarginalize (bool autorange=true)
Marginalize from TTree. More...

void PrepareToContinueMarginalization (const std::string &filename, const std::string &mcmcTreeName="", const std::string &parameterTreeName="", bool loadObservables=true, bool autorange=true)
Continue the marginalization already stored in another file. More...

virtual bool UpdateMultivariateProposalFunctionCovariances (double a)
Update multivariate proposal function covariances. More...

virtual bool UpdateMultivariateProposalFunctionCovariances ()
Update multivariate proposal function covariances. More...

void CalculateCholeskyDecompositions ()
Calculate Cholesky decompositions needed for multivariate proposal function. More...

void UpdateChainIndex (int chain)
Keep track of which chain is currently computed (within a thread). More...

Protected Member Functions
virtual double GraphCorrection (unsigned ibin) const
Take care of bin width when creating a graph from the fit function.

Protected Member Functions inherited from BCIntegrate
virtual std::string GetBestFitSummary (unsigned i) const
Get string summarizing best fit for single variable. More...

unsigned IntegrationOutputFrequency () const
Determine frequency of output during integration.

void LogOutputAtEndOfIntegration (double integral, double absprecision, double relprecision, int nIterations)
Helper method to output at end of integration. More...

void LogOutputAtIntegrationStatusUpdate (BCIntegrationMethod type, double integral, double absprecision, int nIterations)
Helper method to output integration status. More...

void LogOutputAtStartOfIntegration (BCIntegrationMethod type, BCCubaMethod cubatype)
Helper method to output at beginning of integration. More...

virtual void PrintBestFitSummary () const
Print best fit to log.

virtual void PrintMarginalizationSummary () const
Print marginalization to log. More...

Protected Member Functions inherited from BCEngineMCMC
virtual void PrintModelSummary () const
Print model summary to log. More...

void SetFillHistogram (int x, int y, bool flag)
Set whether to fill 2D histogram y vs x: positive indices for parameters; negative for observables, starting at -1 and going more negative—observable index = -(index+1). More...

unsigned UpdateFrequency (unsigned N) const
return appropriate update interval More...

void UpdateParameterTree ()
Update Paramater TTree with scales and efficiencies. More...

Protected Attributes
TH1D fHistogram
The histogram with observed number of counts. More...

Protected Attributes inherited from BCFitter
bool fErrorBandContinuous
A flag for single point evaluation of the error "band".

double fErrorBandExtensionLowEdgeX
extends the lower edge of x range by the given value

double fErrorBandExtensionLowEdgeY
extends the upper edge of y range by the given value

double fErrorBandExtensionUpEdgeX
extends the upper edge of x range by the given value

double fErrorBandExtensionUpEdgeY
extends the upper edge of y range by the given value

unsigned fErrorBandNbinsX
Number of X bins of the error band histogram.

unsigned fErrorBandNbinsY
Number of Y bins of the error band histogram.

std::vector< double > fErrorBandX
The x positions where the error is calculated.

TH2D fErrorBandXY
The error band histogram.

int fFitFunctionIndexX
The index for function fits in x direction.

int fFitFunctionIndexY
The index for function fits in y direction.

BCDataSet fFitterDataSet
Needed for uncertainty propagation.

bool fFlagFillErrorBand
Flag whether or not to fill the error band.

bool fFlagIntegration
Flag for using the ROOT TH1::Integral method (true), or linear interpolation (false)

double fPValue
p value for goodness of fit

Protected Attributes inherited from BCModel
BCH1D fBCH1DPosteriorDrawingOptions
knowledge update plot 1D posterior options. More...

BCH1D fBCH1DPriorDrawingOptions
knowledge update plot 1D prior options. More...

BCH2D fBCH2DPosteriorDrawingOptions
knowledge update plot 2D posterior options. More...

BCH2D fBCH2DPriorDrawingOptions
knowledge update plot 2D prior options. More...

BCDataSet * fDataSet
A data set. More...

bool fDrawPriorFirst
flag for ordering of drawing of prior and posterior in knowledge update plots. More...

bool fFactorizedPrior
flag for whether factorized prior has been used. More...

BCPriorModel * fPriorModel
BCPriorModel object for drawing of knowledge update, and saving of samples according to prior. More...

Protected Attributes inherited from BCIntegrate
bool fFlagIgnorePrevOptimization
Flag for ignoring older results of optimization.

bool fFlagMarginalized
flag indicating if the model was marginalized

double fSALogProb
Log probability of current simulated annealing iteration. More...

int fSANIterations
Number of iterations for simualted annealing. More...

double fSAT0
Starting temperature for Simulated Annealing.

double fSATemperature
Current temperature of simulated annealing algorithm. More...

double fSATmin
Minimal/Threshold temperature for Simulated Annealing.

std::vector< double > fSAx
Current simulated annealing parameter point. More...

Protected Attributes inherited from BCEngineMCMC
BCH1D fBCH1DdrawingOptions
A BCH1D (with no histogram) for storing BCH1D drawing options. More...

BCH2D fBCH2DdrawingOptions
A BCH2D (with no histogram) for storing BCH2D drawing options. More...

bool fCorrectRValueForSamplingVariability
flag for correcting R value for initial sampling variability. More...

std::vector< TH1 * > fH1Marginalized
Vector of 1D marginalized distributions.

std::vector< std::vector< TH2 * > > fH2Marginalized
Vector of 2D marginalized distributions. More...

double fHistogramRescalePadding
factor for enlarging range of histograms when rescaling. More...

unsigned fInitialPositionAttemptLimit
Maximum number of attempts to make to set the initial position. More...

BCEngineMCMC::InitialPositionScheme fInitialPositionScheme
Variable which defines the initial position. More...

std::vector< double > fLocalModes
Vector of local modes. More...

int fMCMCCurrentIteration
The current iteration number. More...

double fMCMCEfficiencyMax
The maximum allowed efficiency for MCMC.

double fMCMCEfficiencyMin
The minimum required efficiency for MCMC.

bool fMCMCFlagWriteChainToFile
Flag to write Markov chains to file.

bool fMCMCFlagWritePreRunToFile
Flag to write pre run to file.

std::vector< std::vector< double > > fMCMCInitialPosition
The intial position of each Markov chain. More...

std::vector< double > fMCMCInitialScaleFactors
User-provided initial values of the scale factors of the factorized proposal function. More...

unsigned fMCMCNChains
Number of Markov chains ran in parallel.

int fMCMCNIterationsConvergenceGlobal
Number of iterations needed for all chains to convergence simultaneously.

unsigned fMCMCNIterationsPreRunCheck
Number of iterations between scale adjustments and convergence checks in pre-run. More...

unsigned fMCMCNIterationsPreRunMax
Maximum number of iterations for a Markov chain prerun.

unsigned fMCMCNIterationsPreRunMin
Minimum number of iterations for the pre-run.

unsigned fMCMCNIterationsRun
Number of iterations for a Markov chain run.

unsigned fMCMCNLag
The lag for the Markov Chain.

TFile * fMCMCOutputFile
Output file for for writing MCMC Tree. More...

std::string fMCMCOutputFilename
Output filename for for writing MCMC Tree. More...

std::string fMCMCOutputFileOption
Output file open option for for writing MCMC Tree. More...

BCEngineMCMC::Phase fMCMCPhase
The phase of the run. More...

unsigned fMCMCPreRunCheckClear
Number of iterations between clearing of convergence stats in pre-run. More...

double fMCMCProposalFunctionDof
Degree of freedom of Student's t proposal. More...

std::vector< std::vector< double > > fMCMCProposalFunctionScaleFactor
Scale factors for proposal functions. More...

bool fMCMCProposeMultivariate
Flag for using multivariate proposal function. More...

std::vector< double > fMCMCRValueParameters
The R-values for each parameter.

double fMCMCRValueParametersCriterion
The R-value criterion for convergence of parameters.

double fMCMCScaleFactorLowerLimit
Lower limit for scale factors.

double fMCMCScaleFactorUpperLimit
Upper limit for scale factors.

std::vector< ChainState > fMCMCStates
The current states of each Markov chain. More...

std::vector< BCEngineMCMC::Statistics > fMCMCStatistics
Statistics for each Markov chain. More...

BCEngineMCMC::Statistics fMCMCStatistics_AllChains
Statistics across all Markov chains. More...

TTree * fMCMCTree
The tree containing the Markov chains. More...

unsigned int fMCMCTree_Chain
Chain number for storing into tree.

ChainState fMCMCTree_State
MC state object for storing into tree.

bool fMCMCTreeLoaded
flag for whether MCMC Tree successfully loaded. More...

bool fMCMCTreeReuseObservables
flag for whether to reuse MCMC Tree's observables. More...

double fMultivariateCovarianceUpdateLambda
weighting parameter for multivariate-proposal-function covariance update. More...

unsigned fMultivariateCovarianceUpdates
Number of multivariate-proposal-function covariance updates performed. More...

double fMultivariateEpsilon
multivariate-proposal-function cholesky-decomposition nudge. More...

std::vector< TMatrixD > fMultivariateProposalFunctionCholeskyDecomposition
Cholesky decompositions for multivariate proposal function. More...

std::vector< TMatrixDSym > fMultivariateProposalFunctionCovariance
Covariance matrices used in multivariate proposal functions. More...

double fMultivariateScaleMultiplier
factor to multiply or divide scale factors by in adjusting multivariate-proposal-function scales. More...

std::string fName
Name of the engine. More...

BCAux::BCTrash< TObject > fObjectTrash
Storage for plot objects with proper clean-up.

BCObservableSet fObservables
User-calculated Observables Set.

BCParameterSet fParameters
Parameter settings.

TTree * fParameterTree
The tree containing the parameter information. More...

TRandom3 fRandom
Random number generator.

std::vector< std::pair< int, int > > fRequestedH2
Vector of pairs of indices for which 2D histograms should be stored. More...

bool fRescaleHistogramRangesAfterPreRun
flag for rescaling of histograms after pre-run. More...

std::string fSafeName
Safe name of the engine for use in naming ROOT objects. More...

Additional Inherited Members
Public Types inherited from BCIntegrate
enum BCOptimizationMethod { kOptEmpty, kOptSimAnn, kOptMetropolis, kOptMinuit, kOptDefault, NOptMethods }
An enumerator for the mode finding algorithm. More...

enum BCIntegrationMethod { kIntEmpty, kIntMonteCarlo, kIntCuba, kIntGrid, kIntLaplace, kIntDefault, NIntMethods }
An enumerator for integration algorithms. More...

enum BCMarginalizationMethod { kMargEmpty, kMargMetropolis, kMargMonteCarlo, kMargGrid, kMargDefault, NMargMethods }
An enumerator for marginalization algorithms. More...

enum BCSASchedule { kSACauchy, kSABoltzmann, kSACustom, NSAMethods }
An enumerator for the Simulated Annealing schedule. More...

enum BCCubaMethod { kCubaVegas, kCubaSuave, kCubaDivonne, kCubaCuhre, kCubaDefault, NCubaMethods }
An enumerator for Cuba integration methods. More...

typedef void(BCIntegrate::* tRandomizer) (std::vector< double > &) const
A pointer for a function that chooses a next random point.

typedef double(BCIntegrate::* tEvaluator) (std::vector< double > &, const std::vector< double > &, bool &)
A pointer for a function that evaluates at a point.

typedef void(* tIntegralUpdater) (const std::vector< double > &, const int &, double &, double &)
A pointer for a function that updates the integral and absolute precision.

Public Types inherited from BCEngineMCMC
enum Precision { kLow, kQuick, kMedium, kHigh, kVeryHigh }
An enumerator for the status of a test. More...

enum Phase { kPreRun = -1, kUnsetPhase = 0, kMainRun = +1 }
An enumerator for the phase of the Markov chain. More...

enum InitialPositionScheme { kInitCenter = 0, kInitRandomUniform = 1, kInitUserDefined = 2, kInitRandomPrior = 3 }
An enumerator for markov-chain position initialization. More...

Static Public Member Functions inherited from BCIntegrate
static void IntegralUpdaterMC (const std::vector< double > &sums, const int &nIterations, double &integral, double &absprecision)
Updates info about integrator.

Static Public Member Functions inherited from BCEngineMCMC
static double RValue (const std::vector< double > &means, const std::vector< double > &variances, unsigned n, bool correctForSamplingVariability=true)
Calculate R value of set of batches of samples—represented by their means and variances, all batches containing the same number of samples—according to Brooks & Gelman, "General Methods for Monitoring Convergence of Iterative Simulations," (1988) More...

Detailed Description

A class for fitting histograms with functions.

Author: Daniel Kollar; Kevin Kröninger

Version: 1.0

Date: 11.2008

This class allows fitting of a TH1D histogram using a TF1 function.

Definition at line 32 of file BCHistogramFitter.h.

Constructor & Destructor Documentation

BCHistogramFitter::BCHistogramFitter	(	const TH1 &	hist,
		const TF1 &	func,
		const std::string &	name = `"histogram_fitter_model"`
	)

Constructor.

Parameters

hist	Histogram with observed number of counts. Conceptually, each entry is an integer. But we don't restrict to the integer data type because one might have a number of entries larger than 2^32.
func	fit function
name	name of the model

Definition at line 30 of file BCHistogramFitter.cxx.

virtual BCHistogramFitter::~BCHistogramFitter ( )

inlinevirtual

The default destructor.

Definition at line 50 of file BCHistogramFitter.h.

Member Function Documentation

double BCHistogramFitter::CalculatePValueFast	(	const std::vector< double > &	par,
		unsigned	nIterations = `100000`
	)

Calculate the p-value using fast-MCMC and the likelihood as test statistic.

The method is explained in the appendix of http://arxiv.org/abs/1011.1674

Parameters

par	The parameter values for which expectations are computed
nIterations	number of pseudo experiments generated by the Markov chain

Returns: p-value

Definition at line 155 of file BCHistogramFitter.cxx.

double BCHistogramFitter::CalculatePValueLeastSquares	(	const std::vector< double > &	par,
		bool	weightExpect = `true`
	)

Calculate the p-value using approximate chi^2 distribution of squared difference for conventional weights.

Approximation is valid for bin contents >5 and not as as good for little data as CalculatePValueLikelihood(), see eq. (32.13), PDG: Statistics, Monte Carlo, Group Theory. Physics Letters B 667, 316-339(2008).

Parameters

par	The set of parameter values used in the model, usually the best fit parameters
weightExpect	use the variance from the expected number of counts (true) or the measured counts (false)

Returns: p-value

Definition at line 206 of file BCHistogramFitter.cxx.

double BCHistogramFitter::CalculatePValueLikelihood ( const std::vector< double > & par )

Calculate the p-value using approximate chi^2 distribution of scaled likelihood.

Approximation is valid for bin contents >5, see eq. (32.12), PDG: Statistics, Monte Carlo, Group Theory. Physics Letters B 667, 316-339(2008).

Parameters

par	The set of parameter values used in the model, usually the best fit parameters

Returns: p-value

Definition at line 177 of file BCHistogramFitter.cxx.

void BCHistogramFitter::DrawFit	(	const std::string &	options = `"HIST"`,
		bool	flaglegend = `false`
	)

virtual

Draw the fit in the current pad.

Implements BCFitter.

Definition at line 109 of file BCHistogramFitter.cxx.

void BCHistogramFitter::Fit ( )

virtual

Performs the fit.

Returns: Success of action.

Implements BCFitter.

Definition at line 88 of file BCHistogramFitter.cxx.

const TH1& BCHistogramFitter::GetHistogram ( )

inline

Returns: The data histogram

Definition at line 59 of file BCHistogramFitter.h.

double BCHistogramFitter::LogLikelihood ( const std::vector< double > & parameters )

virtual

The log of the conditional probability.

Overloaded from BCModel.

Parameters

parameters A vector of doubles containing the parameter values.

Implements BCModel.

Definition at line 65 of file BCHistogramFitter.cxx.

Member Data Documentation

TH1D BCHistogramFitter::fHistogram

protected

The histogram with observed number of counts.

Definition at line 113 of file BCHistogramFitter.h.

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

/root/bat/models/base/BCHistogramFitter.h
/root/bat/models/base/BCHistogramFitter.cxx

Search Results

Public Member Functions

Protected Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation