Logo ROOT  
Reference Guide
 
Loading...
Searching...
No Matches
MCMCCalculator.cxx
Go to the documentation of this file.
1// @(#)root/roostats:$Id$
2// Authors: Kevin Belasco 17/06/2009
3// Authors: Kyle Cranmer 17/06/2009
4/*************************************************************************
5 * Copyright (C) 1995-2008, Rene Brun and Fons Rademakers. *
6 * All rights reserved. *
7 * *
8 * For the licensing terms see $ROOTSYS/LICENSE. *
9 * For the list of contributors see $ROOTSYS/README/CREDITS. *
10 *************************************************************************/
11
12
13/** \class RooStats::MCMCCalculator
14 \ingroup Roostats
15
16 Bayesian Calculator estimating an interval or a credible region using the
17 Markov-Chain Monte Carlo method to integrate the likelihood function with the
18 prior to obtain the posterior function.
19
20 By using the Markov-Chain Monte Carlo methods this calculator can work with
21 model which require the integration of a large number of parameters.
22
23 MCMCCalculator is a concrete implementation of IntervalCalculator. It uses a
24 MetropolisHastings object to construct a Markov Chain of data points in the
25 parameter space. From this Markov Chain, this class can generate a
26 MCMCInterval as per user specification.
27
28 The interface allows one to pass the model, data, and parameters via a
29 workspace and then specify them with names.
30
31 After configuring the calculator, one only needs to ask GetInterval(), which
32 will return an ConfInterval (MCMCInterval in this case).
33 */
34
35#include "Rtypes.h"
36#include "RooGlobalFunc.h"
37#include "RooAbsReal.h"
38#include "RooArgSet.h"
39#include "RooArgList.h"
46#include "TIterator.h"
49#include "RooProdPdf.h"
50
52
53using namespace RooFit;
54using namespace RooStats;
55using namespace std;
56
57////////////////////////////////////////////////////////////////////////////////
58/// default constructor
59
61 fPropFunc(0),
62 fPdf(0),
63 fPriorPdf(0),
64 fData(0),
65 fAxes(0)
66{
67 fNumIters = 0;
69 fNumBins = 0;
72 fSize = -1;
74 fLeftSideTF = -1;
75 fEpsilon = -1;
76 fDelta = -1;
77}
78
79////////////////////////////////////////////////////////////////////////////////
80/// constructor from a Model Config with a basic settings package configured
81/// by SetupBasicUsage()
82
84 fPropFunc(0),
85 fData(&data),
86 fAxes(0)
87{
88 SetModel(model);
90}
91
93 // set the model
94 fPdf = model.GetPdf();
95 fPriorPdf = model.GetPriorPdf();
100 if (model.GetParametersOfInterest())
102 if (model.GetNuisanceParameters())
104 if (model.GetConditionalObservables())
106 if (model.GetGlobalObservables())
107 fGlobalObs.add( *(model.GetGlobalObservables() ) );
108
109}
110
111////////////////////////////////////////////////////////////////////////////////
112/// Constructor for automatic configuration with basic settings. Uses a
113/// UniformProposal, 10,000 iterations, 40 burn in steps, 50 bins for each
114/// RooRealVar, determines interval by histogram. Finds a 95% confidence
115/// interval.
116
118{
119 fPropFunc = 0;
120 fNumIters = 10000;
121 fNumBurnInSteps = 40;
122 fNumBins = 50;
125 SetTestSize(0.05);
127 fLeftSideTF = -1;
128 fEpsilon = -1;
129 fDelta = -1;
130}
131
132////////////////////////////////////////////////////////////////////////////////
133
135{
136 if (a < 0 || a > 1) {
137 coutE(InputArguments) << "MCMCCalculator::SetLeftSideTailFraction: "
138 << "Fraction must be in the range [0, 1]. "
139 << a << "is not allowed." << endl;
140 return;
141 }
142
143 fLeftSideTF = a;
145}
146
147////////////////////////////////////////////////////////////////////////////////
148/// Main interface to get a RooStats::ConfInterval.
149
151{
152
153 if (!fData || !fPdf ) return 0;
154 if (fPOI.getSize() == 0) return 0;
155
156 if (fSize < 0) {
157 coutE(InputArguments) << "MCMCCalculator::GetInterval: "
158 << "Test size/Confidence level not set. Returning NULL." << endl;
159 return NULL;
160 }
161
162 // if a proposal function has not been specified create a default one
163 bool useDefaultPropFunc = (fPropFunc == 0);
164 bool usePriorPdf = (fPriorPdf != 0);
165 if (useDefaultPropFunc) fPropFunc = new UniformProposal();
166
167 // if prior is given create product
168 RooAbsPdf * prodPdf = fPdf;
169 if (usePriorPdf) {
170 TString prodName = TString("product_") + TString(fPdf->GetName()) + TString("_") + TString(fPriorPdf->GetName() );
171 prodPdf = new RooProdPdf(prodName,prodName,RooArgList(*fPdf,*fPriorPdf) );
172 }
173
174 RooArgSet* constrainedParams = prodPdf->getParameters(*fData);
176 delete constrainedParams;
177
178 RooArgSet* params = nll->getParameters(*fData);
180 if (fNumBins > 0) {
181 SetBins(*params, fNumBins);
183 if (dynamic_cast<PdfProposal*>(fPropFunc)) {
184 RooArgSet* proposalVars = ((PdfProposal*)fPropFunc)->GetPdf()->
185 getParameters((RooAbsData*)NULL);
186 SetBins(*proposalVars, fNumBins);
187 }
188 }
189
191 mh.SetFunction(*nll);
194 mh.SetParameters(*params);
198
199 MarkovChain* chain = mh.ConstructChain();
200
201 TString name = TString("MCMCInterval_") + TString(GetName() );
202 MCMCInterval* interval = new MCMCInterval(name, fPOI, *chain);
203 if (fAxes != NULL)
204 interval->SetAxes(*fAxes);
205 if (fNumBurnInSteps > 0)
207 interval->SetUseKeys(fUseKeys);
212 if (fEpsilon >= 0)
213 interval->SetEpsilon(fEpsilon);
214 if (fDelta >= 0)
215 interval->SetDelta(fDelta);
216 interval->SetConfidenceLevel(1.0 - fSize);
217
218 if (useDefaultPropFunc) delete fPropFunc;
219 if (usePriorPdf) delete prodPdf;
220 delete nll;
221 delete params;
222
223 return interval;
224}
#define a(i)
Definition RSha256.hxx:99
#define coutE(a)
const Bool_t kFALSE
Definition RtypesCore.h:92
#define ClassImp(name)
Definition Rtypes.h:364
char name[80]
Definition TGX11.cxx:110
RooArgSet * getParameters(const RooAbsData *data, bool stripDisconnected=true) const
Create a list of leaf nodes in the arg tree starting with ourself as top node that don't match any of...
virtual void removeAll()
Remove all arguments from our set, deleting them if we own them.
Int_t getSize() const
RooAbsData is the common abstract base class for binned and unbinned datasets.
Definition RooAbsData.h:49
virtual RooAbsReal * createNLL(RooAbsData &data, const RooLinkedList &cmdList)
Construct representation of -log(L) of PDFwith given dataset.
RooAbsReal is the common abstract base class for objects that represent a real value and implements f...
Definition RooAbsReal.h:61
RooArgList is a container object that can hold multiple RooAbsArg objects.
Definition RooArgList.h:21
RooArgSet is a container object that can hold multiple RooAbsArg objects.
Definition RooArgSet.h:29
Bool_t add(const RooAbsArg &var, Bool_t silent=kFALSE) override
Add element to non-owning set.
RooProdPdf is an efficient implementation of a product of PDFs of the form.
Definition RooProdPdf.h:37
Bayesian Calculator estimating an interval or a credible region using the Markov-Chain Monte Carlo me...
enum MCMCInterval::IntervalType fIntervalType
virtual MCMCInterval * GetInterval() const
Main interface to get a ConfInterval.
void SetBins(const RooAbsCollection &coll, Int_t numBins) const
void SetupBasicUsage()
Constructor for automatic configuration with basic settings.
MCMCCalculator()
default constructor
virtual void SetLeftSideTailFraction(Double_t a)
Set the left side tail fraction.
virtual void SetTestSize(Double_t size)
set the size of the test (rate of Type I error) ( Eg. 0.05 for a 95% Confidence Interval)
virtual void SetModel(const ModelConfig &model)
Set the Model.
ProposalFunction * fPropFunc
MCMCInterval is a concrete implementation of the RooStats::ConfInterval interface.
virtual void SetAxes(RooArgList &axes)
Set which parameters go on which axis.
virtual void SetNumBurnInSteps(Int_t numBurnInSteps)
set the number of steps in the chain to discard as burn-in, starting from the first
virtual void SetDelta(Double_t delta)
kbelasco: The inner-workings of the class really should not be exposed like this in a comment,...
virtual void SetLeftSideTailFraction(Double_t a)
set the left-side tail fraction for a tail-fraction interval
virtual void SetConfidenceLevel(Double_t cl)
set the desired confidence level (see GetActualConfidenceLevel()) Note: calling this function trigger...
virtual void SetIntervalType(enum IntervalType intervalType)
Set the type of interval to find.
virtual void SetEpsilon(Double_t epsilon)
set the acceptable level or error for Keys interval determination
virtual void SetUseKeys(Bool_t useKeys)
set whether to use kernel estimation to determine the interval
virtual void SetUseSparseHist(Bool_t useSparseHist)
set whether to use a sparse histogram.
Stores the steps in a Markov Chain of points.
Definition MarkovChain.h:30
This class uses the Metropolis-Hastings algorithm to construct a Markov Chain of data points using Mo...
virtual void SetProposalFunction(ProposalFunction &proposalFunction)
virtual void SetChainParameters(const RooArgSet &set)
virtual void SetParameters(const RooArgSet &set)
virtual MarkovChain * ConstructChain()
virtual void SetNumIters(Int_t numIters)
virtual void SetType(enum FunctionType type)
virtual void SetFunction(RooAbsReal &function)
virtual void SetSign(enum FunctionSign sign)
ModelConfig is a simple class that holds configuration information specifying how a model should be u...
Definition ModelConfig.h:30
const RooArgSet * GetConditionalObservables() const
get RooArgSet for conditional observables (return NULL if not existing)
const RooArgSet * GetGlobalObservables() const
get RooArgSet for global observables (return NULL if not existing)
const RooArgSet * GetParametersOfInterest() const
get RooArgSet containing the parameter of interest (return NULL if not existing)
const RooArgSet * GetNuisanceParameters() const
get RooArgSet containing the nuisance parameters (return NULL if not existing)
RooAbsPdf * GetPdf() const
get model PDF (return NULL if pdf has not been specified or does not exist)
RooAbsPdf * GetPriorPdf() const
get parameters prior pdf (return NULL if not existing)
PdfProposal is a concrete implementation of the ProposalFunction interface.
Definition PdfProposal.h:30
UniformProposal is a concrete implementation of the ProposalFunction interface for use with a Markov ...
virtual const char * GetName() const
Returns name of object.
Definition TNamed.h:47
Basic string class.
Definition TString.h:136
RooCmdArg Constrain(const RooArgSet &params)
RooCmdArg GlobalObservables(const RooArgSet &globs)
RooCmdArg ConditionalObservables(const RooArgSet &set)
The namespace RooFit contains mostly switches that change the behaviour of functions of PDFs (or othe...
@ InputArguments
Namespace for the RooStats classes.
Definition Asimov.h:19
void RemoveConstantParameters(RooArgSet *set)