Negative weights in splot

Hello,

I have performed a 2D fit in m(K^{+}K^{-}\mu^{+}\mu^{-}) and m(K^{+}K^{-}), the fit is converging and for reference, the fit results are mentioned in [1]. I need to calculate the sWeights and apply them to data to compare various kinematic variable distributions with MC.

While calculating the sWeights I’m getting negative weights for ~30% of the events. I’m not able to figure out how to avoid getting negative weights. In the past (when fitting over different data samples), the -ive weights were ~5-6% and were ignored but in this case, ~30% may lead to bias in kinematic variable distributions. Please suggest how to avoid this issue.

I’m also attaching the 1D projections of m(K^{+}K^{-}\mu^{+}\mu^{-}) and m(K^{+}K^{-}) data fits where the signal and different background components are mentioned
BmassDataFit_BsToJpsif2p_1Dprojection_bdtr_.pdf (24.8 KB)
PhimassDataFit_BsToJpsif2p_1Dprojection_bdtr_.pdf (23.5 KB)

along with the script used to calculate sWeights

splot.cc (10.4 KB)

root -l -b -q 'splot.cc(0, 2016)'

the RooWorkspace used as input to the scripts
ws_BsToJpsif2p_2DfitResults_2016.root (338.0 KB)

and the lxplus path (public) of the dataset

/afs/cern.ch/work/r/rraturi/public/sWeight

Please let me know if anymore information is required from my side, thanks!

Best regards,
Rishabh

[1]

  RooFitResult: minimized FCN value: -271219, estimated distance to minimum: 0.000674602
                covariance matrix quality: Full, accurate covariance matrix
                Status : MINIMIZE=0 HESSE=0 MINOS=0 

    Constant Parameter    Value     
  --------------------  ------------
  alpha1_BdToJpsiKpi_2016    3.0000e+00
  alpha1_BsToJpsiKK_2016    2.4254e+00
  alpha1_BsToJpsif2p_2016    2.9267e+00
  alpha1_LambdaBToJpsiKp_2016    3.1811e-02
  alpha2_BdToJpsiKpi_2016   -5.3222e-01
  alpha2_BsToJpsiKK_2016   -1.9378e+00
  alpha2_BsToJpsif2p_2016   -2.1155e+00
  decaywidth_BsToJpsif2p_2016    7.6827e-02
  frac_BdToJpsiKpi_2016    2.7284e-01
  frac_BsToJpsiKK_2016    4.3841e-01
  frac_BsToJpsif2p_2016    4.7349e-01
  mean_BsToJpsiKK_2016    5.3663e+00
  mean_LambdaBToJpsiKp_2016    5.3747e+00
   n1_BdToJpsiKpi_2016    5.0000e-01
    n1_BsToJpsiKK_2016    5.0000e-01
   n1_BsToJpsif2p_2016    5.0000e-01
  n1_LambdaBToJpsiKp_2016    5.0000e-01
   n2_BdToJpsiKpi_2016    1.0000e+01
    n2_BsToJpsiKK_2016    1.0000e+01
   n2_BsToJpsif2p_2016    1.0000e+01
  sigma1_BdToJpsiKpi_2016    6.6174e-02
  sigma1_BsToJpsiKK_2016    2.3456e-02
  sigma1_BsToJpsif2p_2016    2.3718e-02
  sigma1_LambdaBToJpsiKp_2016    6.5791e-02
  sigma2_BdToJpsiKpi_2016    4.5359e-02
  sigma2_BsToJpsiKK_2016    4.0515e-02
  sigma2_BsToJpsif2p_2016    4.1995e-02
  slope_BdToJpsiKpi_2016    2.2828e-02
  slope_BsToJpsiKK_2016    9.2544e-03

    Floating Parameter  InitialValue    FinalValue (+HiError,-LoError)    GblCorr.
  --------------------  ------------  ----------------------------------  --------
  mean_BdToJpsiKpi_2016    5.4200e+00    5.4284e+00 (+3.15e-03,-3.21e-03)  <none>
  mean_BsToJpsif2p_2016    5.3670e+00    5.3671e+00 (+1.01e-04,-9.91e-05)  <none>
  mean_relBW_BsToJpsif2p_2016    1.5250e+00    1.5234e+00 (+8.74e-04,-8.71e-04)  <none>
  nbkgE_BsToJpsif2p_2016    5.8000e+03    4.0073e+03 (+6.86e+02,-6.76e+02)  <none>
  nsigD_BsToJpsif2p_2016    3.5000e+03    3.5856e+03 (+2.59e+02,-2.57e+02)  <none>
  slope_BsToJpsif2p_2016   -5.0000e+00   -1.5073e+00 (+1.30e+00,-1.13e+00)  <none>
  slopef2p_BsToJpsif2p_2016    1.0000e-01    7.0559e-10 (+5.00e-03,--0.00e+00)  <none>
   yc_BdToJpsiKpi_2016    1.4880e+01    1.7211e+00 (+1.67e-01,-1.43e-01)  <none>
    yc_BsToJpsiKK_2016    3.2000e+00    1.1580e+00 (+1.80e-01,-1.57e-01)  <none>
  yc_LambdaBToJpsiKp_2016    1.5100e+00    1.2184e+00 (+2.68e-01,-2.89e-01)  <none>

Hi @RISHABH_RATURI1,

Thank you for your post.
@jonas could you please have a look?

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.