---

After discussion on - 06.05.2015

---

rc clean & rc clean_dep - solve the problem with recompillation and vertioning of different tau and muon identification files.

Now the muons and tau jets can be read out in a simple analysis file --> Plot the pt distributions. How to find MET?

---

16.04.2015 Discussion with Lorenz

---

here are a few more information on how to do the analysis.

1) Markus' thesis: https://cds.cern.ch/record/1694337

2) https://twiki.cern.ch/twiki/bin/viewauth/AtlasComputing/SoftwareTutorialxAODAnalysisInROOT It probably makes sense to go through this (some points are not very important for the beginning I guess. Maybe it would be best for the beginning to get the taus from the xAOD and fill a histogramm with the pt, eta, phi etc.)

3) there's a xAOD mini tutorial: https://indico.cern.ch/event/362819/other-view?view=standard unfortunately there is no real documentation of the hands-on part, maybe they also just went through the TWiki above

4) If you need a sample at our institute, I have one here: /ZIH.fast/users/hauswald/samples_HIGG4D1_p1851

The RunII Analysis Model

• GoThroughXAODanalysis

---

24.07.2014

---

How to run python scripts on lxplus to get right files:

RunPythonScriptsOnLXPLUS

---

19-20.05.2014

---

I found that supercells contain tile if to decode it correctly, but do not get the layer 1.

I assume, that electromagnetic fraction of the electrons will try to go to 1, while for jets it should fluctuate and be more uniform, while for taus it should be more similar to electrons but more broad and then one can use it for the separation of taus from jets.

What do I need to separate from what?

Now it makes FUN finally!

---

---

Think I do not understand is - delta R calculations - is it with eta and phy numbers or values?

Looks like I did it with numbers and it can be not fully correct!

---

On DESY laptop:

/home/daq/Desktop/first_sample/test_ntuple/test

make

./Analysis - works - one of the vertions from the pktlogin

It did not work due to some strange effects on the pktlogin - due to increneting "i" in a for loop with fixed maximum. some changes happened?

I put the EM_to_HAD variable, but it looks like trig_L1_emtau_hadCore->size() = 67 & trig_L1_emtau_n->size() = 67 0 ::: 0 1 ::: 0 2 ::: 1000 3 ::: 0 4 ::: 0 5 ::: 0 6 ::: 0 7 ::: 1000 8 ::: 0 9 ::: 0 10 ::: 0 11 ::: 0 12 ::: 0 13 ::: 0 14 ::: 0 15 ::: 0 16 ::: 0 17 ::: 0 18 ::: 0 19 ::: 0 20 ::: 0 21 ::: 0 22 ::: 0 23 ::: 2000 24 ::: 0 25 ::: 33000 26 ::: 0 27 ::: 0 28 ::: 0 29 ::: 0 30 ::: 0 31 ::: 0 32 ::: 0 33 ::: 0 34 ::: 2000 35 ::: 0 36 ::: 0 37 ::: 0 38 ::: 0 39 ::: 0 40 ::: 0 41 ::: 0 42 ::: 0 43 ::: 0 44 ::: 0 45 ::: 0 46 ::: 0 47 ::: 0 48 ::: 0 49 ::: 0 50 ::: 0 51 ::: 0 52 ::: 0 53 ::: 2000 54 ::: 0 55 ::: 0 56 ::: 0 57 ::: 1000 58 ::: 0 59 ::: 3000 60 ::: 0 61 ::: 1000 62 ::: 0 63 ::: 0 64 ::: 3000 65 ::: 3000 66 ::: 0

All data files from the pktlogin I copied to transcend media driver to process with DESY laptop

Main issue to get variable working as ratio between hadronic and electromagnetic showers and see difference between tau and jets and electrons.

From what I remember, we were using trig_L1 objects for the analysis. For example, the HadCore variable was directly coming from trig_L1_emtau_hadCore. The had. isolation is also available through trig_L1_emtau_hadIsol. With this kind of objects, you should be able to use what you want. Some informations are, of course, not available at L1, you can either recompute them or look at L2 or EF variables.

if the trigger tower works it should be fine. But I still think the Tile info is somehow in the super cell list.

---

Tau studies including hadronic information

---

http://hep.phys.sfu.ca/theses/MTMcDonald_msc.pdf

https://twiki.cern.ch/twiki/bin/view/LAr/PhaseITDRSimulation

I found

vector *L1CaloPPM_hadTTCellsEnergy;

vector *L1CaloPPM_hadTTCellsEt;

vector *L1CaloPPM_hadTileCellsEnergy;

vector *L1CaloPPM_hadCoolId;

vector *L1CaloPPM_hadTTCellsEnergy;

vector *L1CaloPPM_hadTTCellsEt;

vector *L1CaloPPM_hadTileCellsEnergy;

vector<vector > *L1CaloPPM_hadTTCellsEnergyByLayer;

vector<vector > *L1CaloPPM_hadTTCellsEtByLayer;

vector<vector > *L1CaloPPM_hadTTCellsEnergyByReceiver;

vector<vector > *L1CaloPPM_hadTTCellsEtByReceiver;

vector *L1CaloPPM_hadPeak;

Int_t tower_n;

vector *tower_tower_pT;

vector *tower_tower_eta;

vector *tower_tower_phi;

vector *tower_tower_E;

vector *tower_tower_nCaloCells;

vector<vector > *tower_tower_CellWeights;

vector *tower_tower_statusWord;

vector *L1CaloPPM_emCoolId;

vector *L1CaloPPM_emTTCellsEnergy;

vector *L1CaloPPM_emTTCellsEt;

vector<vector > *L1CaloPPM_emTTCellsEnergyByLayer;

vector<vector > *L1CaloPPM_emTTCellsEtByLayer;

vector<vector > *L1CaloPPM_emTTCellsEnergyByReceiver;

vector<vector > *L1CaloPPM_emTTCellsEtByReceiver;

vector *L1CaloPPM_emPeak;

vector *L1CaloPPM_emADCPeak;

vector<vector > *L1CaloPPM_emLUT;

vector<vector > *L1CaloPPM_emADC;

vector *L1CaloPPM_emEnergy;

The goal -- to prove readout of triggering system working and working efficient, stable, correctly.

For this AREUS simulation program is under development - https://cds.cern.ch/record/1546755

Several options should be done - optimisation of the filter parameters - batchelor students.

TILE Calorimeters implementation - and running AREUS for fcore variables.

Fcore variables final optimisation and improvements.

FCAL calorimeter implementation in AREUS - for the filter studies for FCAL and later verification of the AREUS and ATHENA... ( possible tast for PhD student)

Changed dimintions of result and efficiency histograms for REM now it is from 0 to 1, before was 0 to 10. What is stupid.

Need to test rejection of variables with respect of the size of area --> as bigger objects might be found with lower rate then smaller and this can create effect on the rate.

Found mistake in REM variable histogramms ( binning) and changed it to have appropriate binning to get rates.

Produced plot of BDT efficiencies with respect to HA11 and tau mc_truth information

BDT flag has to go and only be used to compare results of analysis to the offline BDT reconstruction of taus. I need to calculate number of all taus, number of accepted taus at different fcore variables and how many are accepted by BDT flag and what is an overlap.

tau_JetBDTSigMedium - In offline analysis, the BDT is used for tau id.

In addition, the likelihood and BDT are trained for different numbers of reconstructed vertices in order to take event pile-up into account. Three dedicated working points with signal efficiencies of ~60%,~45% and ~30% (loose, medium, tight) are provided for all tau identification methods.

TDR --> https://cds.cern.ch/record/1597130/files/ATL-COM-GEN-2013-007.pdf

Changed number of varied scells to 6

Added R_EM variable - test it now to see the distribution.

Plotted Efficiency for the f_core_raw variable F_core_rnu efficiencies look not convincing.

Use HA11 for making plots for rate as a function of pt in the second part of events

Efficiency after tau truth ad bdt --> HA11 as a number

---

How to connect to the iktp batch system:

---

>ssh -XYC olga@pktloginNOSPAMPLEASE.phy.tu-dresden.de

>ssh -XYC pktf45

>module load root

/home/olga/IKTP/TauAnalysis/RWC_ATLASntuples - folder with Robert code to change for the new ntuple CaloD3PDh CaloD3PDC

---

21.08.2013 Tau analysis

---

I can create 25 histograms for all variables as number of bins this will emitate the threshold for the variables and for each bin of variable I can set each ROI independent if it is a jet or electron or tau to the histogram of Pt vs Number of events times weight. Later this plots can be analyzed by the script and integrate - everything what is above a certain threshold is rate.

Histograms for rates and then integral over them are created.

I have to make more backups!

---

20.08.2013 Tau analysis

---

What is done:

There are studies of several variables with likelihood analysis.

1. For the half of data the variables are calculated and the distributions take into account the mc_event_weight. For studies jet and taus tau_eta/phi branch is used, for the jet searching jet_antikt4topo_em_eta/phi is used. Around eta and phi hot super cell (scell) is found and the variables are calculated from this hot scell.

2. After this distributions are collected the normalized plots are calculated.

3. From the normalized values the likelihoods values are calculated for the tau to jet and jet to tau:

L_tau = V_t / ( V_t + V_j )

L_jet = V_j / ( V_t + V_j )

4. For the next half the events the ROI trig_L1_emtau_eta and phi are taken and are matched to tau or jet. then the variables are calculated and for tau or jet for the calculated variable likelihood is corresponded and using weight is filled in "Result" histograms.

5. "Result" histograms are used after data reading to calculate "Efficiency" histograms and "jjRate" histograms. For efficiency histograms the loop over bins is used and calculated sum over bins then to calculate efficiency:

Eff=(integral-sum_over_bins)/integral

The function from 100% to 0 with increasing bin number should be observed.

6. For the jjRate histogram the rate over jets is calculated from result histograms for different thresholds for likelihoods as a function of efficiency for the tau identification. The plot with increasing rate of the jets is expected with increasing efficiency to the tau identification.

The luminosity and filter efficiency was not used so far. The rate of events for different likelihood cuts as a function of pt is missing.

---

14.08.2013

---

My code is working with new ntuples, I cereated new calod3pd files and with them I can run my code. I do loop over taus and see if they match to truth tau information or to jets.

SOme parts crashing, but most of variables working.

I need to try Robert code with new CaloD3Pd

---

* 13.08.2013 How to improve our studies

---

Interesting to see how efficient are the variables we study, f_core, f_raw, r_nu, ringisolation and so on. We have official ntuples from ATLAS. Arno will download them. Samples https://aagaard.web.cern.ch/aagaard/mc12_14TeV/mc12_NTUP_COMMON.html

link for the dijet samples https://twiki.cern.ch/twiki/bin/viewauth/AtlasProtected/JetStudies2012#Weighting_scheme_and_MC_quality

To do:

look for taus reconstructed as tau jet and see if all our variables allow us to reconstruct the same taus:

vector *tau_jet_Et;

vector *tau_jet_pt;

vector *tau_jet_m;

vector *tau_jet_eta;

vector *tau_jet_phi;

Next vectors to use to find taus:

Int_t tau_n;

vector *tau_Et;

vector *tau_pt;

vector *tau_m;

vector *tau_eta;

vector *tau_phi;

Then to take mc_* information:

Int_t mc_n;

vector *mc_pt;

vector *mc_m;

vector *mc_eta;

vector *mc_phi;

vector *mc_status;

vector *mc_barcode;

vector *mc_pdgId;

And match real tau not decaying to electrons or muons ( hadronically decaying taus) to tau_eta and tau_phi. For them to run variables calculations and compare with previous analysis for other ntuples.

For looking for jets take

Int_t jet_AntiKt4TopoEM_n;

vector *jet_AntiKt4TopoEM_E;

vector *jet_AntiKt4TopoEM_pt;

vector *jet_AntiKt4TopoEM_m;

vector *jet_AntiKt4TopoEM_eta;

vector *jet_AntiKt4TopoEM_phi;

Take into account weighting:

UInt_t mc_event_number;

Float_t mc_event_weight;

vector *tau_jet_flavor_weight;

For mc_... it is just in case I accept tau I just weight it in the histograms.

---

---

Hist_fcore_square_Likelihood[it][iout][i]=new TH1F(name8,title8,nbin_fcore,0,ibin_fcore);

Hist_fcore_square_norm[it][iout][i]=new TH1F(name8,title8,nbin_fcore,0,ibin_fcore);

Hist_fcore_square_Result[it][iout][i]=new TH1F(name8,title8,nbin_fcore,0,ibin_fcore);

Hist_fcore_square_Efficiency[it][iout][i]=new TH1F(name8,title8,nbin_fcore,0,ibin_fcore);

Hist_fcore_square_j[iout][i]=new TH1F(name8,title8,nbin_fcore,0,ibin_fcore);