RooStatsKarlsruhe has been now incorporated in RooStatsCms, a more comprehensive. All the old functionalities are still present but you can find a lot more! Take a look to RooStatsCms site here
RooStatsKarlsruhe (RSK) is a framework for statistical studies based on RooFit ,an advanced toolkit for data modeling which is already part of ROOT. RSK is composed of two main parts:
RooStatsKarlsruhe depends only on ROOT 5.20.00. This version of the package has been developed and tested with Scientific Linux 4 (Intel(R) Xeon(R) CPU 5150, gcc 3.1.4 ) and Debian 4.0 Etch (amd64 gcc 4.1.2). Recently a lot of changes were added to RooFit, also driven by the growth of RooStatsCms. We strongly recomend to use the latest version of roofit (HEAD) to profit from all the improvements. To obtain it, just use this svn command:
- Statistical methods
- Graphical routines It comes with a collection of examples and a rich documentation in doxygen style. During the make the CINT dictionaries are created: you can therefore use the classes of the package interactively with Root or in macros. RSK integragrates an implementation of the constraints different from the one of RooFit, please see the documentation of the Constraint class and our examples .
The package in a tarball is available here!
Be sure that you have the proper ROOT version set up and just copy-paste!(be careful with the version..) The following lines will create a library dynamic called libRooStatsKarlsruhe.so in the lib directory. See the directory macros/examples for some example macros, ready to be run.
svn co -rHEAD http://root.cern.ch/svn/root/trunk/roofit
Setting Root on the lxplus machines is easy and works out of the box just copy and paste these lines (bash shell).
tar -zxf RSK_2.0.0.tar.gz
setenv ROOTSYS /afs/cern.ch/sw/lcg/external/root/5.20.00/slc4_ia32_gcc34/root/
setenv LD_LIBRARY_PATH $ROOTSYS/lib
set path=($path $ROOTSYS/bin)
Also the last line is needed to build the cint dictionaries!
You can find three examples in the macros/examples directory: two for the statistical methods and the third for the LEPBandPlot class.
A very general program, ratio_finder.exe, scans a variable in the model inside a max and a min value so to obtain a desired CLs value. The class used inside the program is RatioFinder. At the end of the procedure the RatioFinderResults object is written to a root file and a png image of the evolution of the scan is dumped. Useful to answer questions like:
- I want to do a green/yellow plot "a la Tevatron" where I exclude a certain R factor on the SM cross section
- I want to know how much lumi I need for a CLs of X percent
- Gregory Schott: gregory.schott<at>cern.ch
- Danilo Piparo: danilo.piparo<at>cern.ch
- Guenter Quast
Generated on Tue Jan 6 15:11:39 2009 for RooStatsKarlsruhe by