Uploaded image for project: 'ROOT'
  1. ROOT
  2. ROOT-8033

Poisson constraint for statistical error in HistFactory

    XMLWordPrintable

    Details

      Description

      Hello, reading the document about HistFactory:

      https://cdsweb.cern.ch/record/1456844/files/CERN-OPEN-2012-016.pdf

      I see in equation 9 that the constraints on the statistical error of the MC are Poissonian. This is not the case when I simply call:

      background.ActivateStatError()
      channel.AddSample(background)
      measurement.AddChannel(channel)

      ws.obj("gamma_stat_channel_bin_1_constraint").Print()
       
      RooGaussian::gamma_stat_channel_bin_1_constraint[ x=nom_gamma_stat_channel_bin_1 mean=gamma_stat_channel_bin_1 sigma=gamma_stat_channel_bin_1_sigma ] = 1

      I guess that this is beacuse of this line, which initialize it to be a gaussian constraints.

      https://root.cern.ch/gitweb?p=root.git;a=blob;f=roofit/histfactory/inc/RooStats/HistFactory/Systematics.h;h=d1000bc3bad896745b6d711582d9b666e23d9fad;hb=HEAD#l359

      Can you:

      1. better document or
      2. use Poissonian constraints as written in the documentation?

        Attachments

          Activity

            People

            • Assignee:
              shageboe Stephan Hageboeck
              Reporter:
              turra Ruggero Turra
            • Votes:
              1 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:
                PlannedEnd:
                PlannedStart: