How to select lnqq events using the Common Ntuples

Introduction

Semileptonic events are selected as mnqq, enqq or tnqq in an exclusive way by applying the different selections one by one and stopping the selection as soon as the event is selected in one of the channels.

The priorities are: 

1. mnqq

2. enqq

3. tnqq (hadronic selection)

4. tnqq (e/m selection)

The priority between mnqq and enqq channels is driven by the lepton identification algorithm which looks for muon, first, and then for electrons.

What follow are the instructions to select the events using the variables which can be found in the CN. If you don't want to loose your time and prefer to read fortran instead of my poor English you can look at the routines in the file count_event_xsec.F in the CN code.

e/mnqq event selection

Event preselection

The e/mnqq events are preselected by requiring that PRELVQQ = 0 and IPRESEM = 1 (electron) or IPRESEM = 2 (muon)

BE CAREFUL: from CN v1.60 IPRESEM can be = 3 for systematic checks, therefore using IPRESEM.GE.1 can be dangerous.

Event selection

The procedure to select electron and muon semileptonic events using the Common Ntuples depends whether you want to apply the latest selection procedure, based on NN functions trained by Anne Ealet, or previous selection procedure, and which CN version you have been analyzing. In addition the following cut has to be applied:

New charged multiplicity cut: NCH.GE.7

NN-based selection function (by Anne Ealet) PRESENT DEFAULT

Common Ntuple version 1.60 or later PREFERRED NTUPLES

The variable SEMSEL is filled with the appropriate (with respect to the lepton flavor and the center of mass energy) NN-based function output. Therefore to select  enqq and mnqq events you have to require SEMSEL.gt.0.6

Common Ntuple between version 1.50 and version 1.511

The variable SEMSEL is filled with  NN-based function output, but the functions are not the latest ones. Therefore to select  enqq and mnqq events you can require SEMSEL.gt.0.6 to have sensible results but not identical to the latest ones or  call the latest version of the NN functions, that you can find in the CN source directory, in your own code and apply the same cut

Common Ntuple version 1.42 or earlier

No variable in the Common Ntuples contains the NN-based function outputs, therefore to select  enqq and mnqq events you can call the latest version of the NN functions, that you can find in the CN source directory, in your own code and apply the same cut as above

PDF-based selection function (aka prob function) (by Paolo Azzurri) OBSOLETE SETTING

Common Ntuple version 1.42 or earlier

The events are selected by requiring that SEMSEL .gt. 0.4 

Common Ntuple version 1.50 or later

Presently it is not possible to apply the PDF-based selection function to this version of the Common Ntuples

tnqq event selection

Event veto

Events which are selected as e/mnqq are rejected

Event preselection

The tnqq events are preselected by requiring that PRELVQQ = 0 and IPRETAU = 1

Event selection

Hadronic selection (follow this link for the NN functions)

• TAUVAR(7).GE.50.AND.TAUVAR(22).GE.5.AND.PTMISS.GE.10.AND.TAUVAR(5).LE.175 where PTMISS=SQRT(PMISS(1)**2+PMISS(2)**2)

• TAUSEL.GT.0.4 if centre of mass energy greater than 186 GeV or TAUSEL.GT.0.6 if centre of mass energy smaller than 186 GeV. 

e/m selection (follow this link for the NN functions)

• (IPRESEM.EQ.1.OR.IPRESEM.EQ.2).AND.NCH.GE.6

• TAUVAR(50).GT.0.4

BE CAREFUL: TAUSEL and TAUVAR(50) contain the correct NN function outputs from CN v1.60. Between version v1.50 and v1.511 they contain outputs of preliminary NN functions. Therefore for CN before v1.60 it is better to call the NN functions in the selection function.

08/10/03