StandardHypoTestInverter

A more sensible implementation of a RooStats demo:

https://root.cern.ch/doc/v608/StandardHypoTestInvDemo_8C.html

Requirements

ROOT v6
RooFit
RooStats
(optional) PROOF
(optional) Boost

Build

Type make to compile. The output should be in bin/. If it fails, make sure you have all the requirements installed. If you don't have Boost installed, you'll have to re-write src/main.cpp yourself and find some other way to set the options. If you still need help, contact me.

In principle you can type make bin/whatever to rename the binary. By default, the output is bin/main.

Input

You must provide a RooWorkSpace inside a .root file. The workspace must contain:

a ModelConfig for signal+background,
a defined observable and parameter of interest,
a RooDataSet containing the data,
(optional) a ModelConfig for background,
(optional) a prior PDF with nuisance parameters.

An example with nuisance parameters would be:

RooWorkspace workspace("workspace");
workspace.import(poi) // Parameter of interest
workspace.import(fitmodel) // Signal+background PDF
workspace.import(data) // Dataset imported on variable x
workspace.import(prior) // Prior PDF
workspace.defineSet("obs",RooArgSet(x));
workspace.defineSet("poi",RooArgSet(poi));
ModelConfig sb_model("sb_model",workspace);
sb_model.SetPdf(*workspace.pdf("fitmodel"));
sb_model.SetPriorPdf(*workspace.pdf("prior"));
sb_model.SetNuisanceParameters(RooArgSet(nuisPar)); // Parameter set contained within prior PDF
sb_model.SetObservables(*workspace.set("obs"));
sb_model.SetParametersOfInterest(*workspace.set("poi"));
workspace.import(sb_model);

The names of the objects are passed to the program at the command line.

Usage

The syntax is bin/main InputFile Workspace ModelSB Data [Options].

Command line option	Description
`--ModelB`	optional background model name
`--NuisPrior`	optional nuisance prior name
`--EnableDetOutput` (default 0)	enable detailed output with all fit information for each toys (output will be written in result file)
`--PlotHypoTestResult` (default 1)	plot test statistic result at each point
`--WriteResult` (default 1)	write HypoTestInverterResult in a file
`--PrintLevel` (default 0)	print level for debugging PL test statistics and calculators
`--ConfLevel` (default 0.95)	confidence level value
`--MassValue`	extra string to tag output file of result
`--ResultFileName`	file with results (by default is built automatically using the workspace input file name)
`--UseCLs` (default 1)	scan for CLs (otherwise for CLs+b)
`--InitialFit` (default -1)	do a first fit to the model (-1 : default, 0 skip fit, 1 do always fit)
`--NPoints` (default 6)	number of points to scan , for autoscan set npoints = -1
`--MaxPoi` (default -1)	max value used of POI (in case of auto scan)
`--NToysRatio` (default 2)	ratio Ntoys(S+B)/Ntoys(B)
`--PoiMin` (default 0)	min/max value to scan in case of fixed scans
`--PoiMax` (default 5)	(if min > max, try to find automatically)
`--UseProof` (default 1)	use Proof Lite when using toys (for freq or hybrid)
`--ReuseAltToys` (default 0)	reuse same toys for alternate hypothesis (if set one gets more stable bands)
`--NToys` (default 1000)	number of toys per point
`--NWorkers` (default 0)	number of worker for ProofLite (default use all available cores)
`--RandomSeed` (default -1)	random seed (if = -1: use default value, if = 0 always random ) NOTE: Proof uses automatically a random seed
`--Rebuild` (default 0)	re-do extra toys for computing expected limits and rebuild test stat distributions (N.B this requires much more CPU (factor is equivalent to nToyToRebuild)
`--NToyToRebuild` (default 100)	number of toys used to rebuild
`--RebuildParamValues` (default 0)	= 0 do a profile of all the parameters on the B (alt snapshot) before performing a rebuild operation (default) = 1 use initial workspace parameters with B snapshot values = 2 use all initial workspace parameters with B Otherwise the rebuild will be performed using [text missing, I assume it works like 2]
`--GenerateBinned` (default 0)	generate binned data sets
`--NoSystematics` (default 0)	force all systematics to be off (i.e. set all nuisance parameters as constant to their nominal values)
`--Optimize` (default 1)	optimize evaluation of test statistic
`--UseNLLOffset` (default 0)	use NLL offset when fitting (this increase stability of fits)
`--UseVectorStore` (default 1)	convert data to use new roofit data store
`--CalculatorType` (default 0)	= 0 Freq calculator = 1 Hybrid calculator = 2 Asymptotic calculator = 3 Asymptotic calculator using nominal Asimov data sets (not using fitted parameter values but nominal ones)
`--NAsimovBins` (default 0)	number of bins in observables used for Asimov data sets (0 is the default and it is given by workspace, typically is 100)
`--TestStatType` (default 2)	= 0 LEP = 1 Tevatron = 2 Profile Likelihood = 3 Profile Likelihood one sided (i.e. = 0 if mu < mu_hat) = 4 Profile Likelihood signed ( pll = -pll if mu < mu_hat) = 5 Max Likelihood Estimate as test statistic = 6 Number of observed event as test statistic
`--UseNumberCounting` (default 0)	set to true when using number counting events
`--MinimizerType`	minimizer type (default is what is in ROOT::Math::MinimizerOptions::DefaultMinimizerType()

Using custom RooFit classes

The folder ExtraRooFit/ is designed to accomodate additional custom RooFit classes. It comes with RooStudentT as an example. In order to add one of your own, place the header (ending in .h) in ExtraRooFit/include/ and the source (ending in .cpp) in ExtraRooFit/src/, then add a line in ExtraRooFit/include/LinkDef.h:

#pragma link C++ class RooCustomPDF+;

where you substitute RooCustomPDF for the appropriate class name. The right file extensions are important, as they need to be picked up in the Makefile.

admorris / StandardHypoTestInverter