Unfolding. A toolbox that connects an experiment s observable variables with true physical quantities. R Matrix connect. Observed distribution THEN

Size: px

Start display at page:

Download "Unfolding. A toolbox that connects an experiment s observable variables with true physical quantities. R Matrix connect. Observed distribution THEN"

Emory Jacobs
5 years ago
Views:

1 Unfolding A toolbox that connects an experiment s observable variables with true physical quantities. Observed distribution R Matrix connect & True distribution THEN True distribution Estimate Inverse R Matrix (M) & Observed distribution!1

2 Unfolding A PyUnfold is used in this work: We need: Crab data is used MC is used to compute it To compute the Flux: It is apply to the unfolding output PyUnfold/FluxFit.py!2

3 Quality cuts rec.anglefitstatus==0 rec.corefitstatus==0 rec.zenithangle<0.785 rec.nchtot>=800 rec.nchavail>0.9*rec.nchtot rec.corefiduscale <= *rec.nChAvail<rec.nHitSP20 rec.nhitsp20<=1.*rec.nchavail!

4 Observed distribution Crab data of 2016 & 2017: /data/archive/hawcroot/data/hawc/data/subsets/crab-strip Apply G/H separation: Sam cuts: /data/scratch/userspace/pretz/scrappy-platypus-dev/optimize- PINCLICompactOptimizer_FHitNNEnergyBinningLogicCFS/cuts-filtered.cuts Kelly cuts: /data/scratch/userspace/kmalone/sweets-for-systematics/nominal/cuts/cuts-fhit-variable.txt Quality cuts: rec.anglefitstatus==0 rec.corefitstatus==0 rec.zenithangle<0.785 rec.nchtot>=800 rec.nchavail>0.9*rec.nchtot rec.corefiduscale <= *rec.nChAvail<rec.nHitSP20!

5 Observed distribution Script used: but it was modified to get all events into a radius from a point Crab: RA: º DEC: 8.62º RA: Crab RA + º DEC: Crab DEC º Subtraction: Observed distribution = Signal + Background (0.5º) - Background (0.5º)!5

6 Observed distribution Frequency GP: all energy bins GP: >1TeV NN: all energy bins NN: >1TeV 2 1 Reco Energy NNV2[GeV] [GeV] 5!6

7 Detector efficiency & response matrix Script used: MC take : /data/scratch/userspace/pretz/daqsim-reconstruction/output/daqsim-baselinetake/gamma.xcd Apply quality cuts & G/H separation: We use the OneWgt because the efficiency and response matrix are normalized.!7

8 Efficiency: Non-Normed Combined Efficiency#quad θ [0.0,5.0) Efficiency GP: all energy bins GP: >1TeV NN: all energy bins NN: >1TeV 5 True Energy [GeV]!8

9 Response Matrix: all & GPEnergyV2 + GP cut!9

10 Response Matrix: all & NNEnergyV2 + NN cut!

11 Response Matrix: >1TeV & GPEnergyV2 + GP cut!11

12 Response Matrix: >1TeV & NNEnergyV2 + NN cut!12

13 Unfold, Final Unfolded Cause Distribution bin0 Frequency 8 7 GP: all energy bins GP: >1TeV NN: all energy bins NN: >1TeV 6 5 July 27th, 2017 Tomás Capistrán Rojas True Energy [GeV] 5!1

14 Flux, script: FluxFit.py Cause Flux Flux 11 GP: all energy bins GP: >1TeV NN: all energy bins NN: >1TeV July 27th, 2017 Tomás Capistrán Rojas True Energy [GeV] 5!1

15 Function fit: N E Y N Y Chi2 / NDF Chi2 NDF NN cut, NN energy all NN cut NN energy >1TeV GP cut GP energy all GP cut GP energy >1TeV 1.5E E E July 27th, 2017 Tomás Capistrán Rojas!15

16 E^2 flux, July 27th, 2017 Tomás Capistrán Rojas!16

17 E^2 flux, July 27th, 2017 Tomás Capistrán Rojas!17

18 Backslide: July 27th, 2017 Tomás Capistrán Rojas!18

19 Obs: GPEnergyV2 + GP cuts EnergyGPV2 (Signal) Frequency GP: all energy bins GP: >1TeV 2 Reco Energy GPV2[GeV] [GeV] 5!19

20 Obs: NNEnergyV2 + NNcuts EnergyNNV2 (Signal) Frequency NN: all energy bins NN: >1TeV 2 1 Reco Energy NNV2[GeV] [GeV] 5!20

21 Efficiency: GPEnergyV2 + GP cuts Non-Normed Combined Efficiency#quad θ [0.0,5.0) Efficiency GP: all energy bins GP: >1TeV 5 True Energy [GeV]!21

22 Efficiency: NNEnergyV2 + NN cuts Non-Normed Combined Efficiency#quad θ [0.0,5.0) Efficiency NN: all energy bins NN: >1TeV 5 True Energy [GeV]!22

23 Unfold, GPEnergyV2 + GP cut Final Unfolded Cause Distribution bin0 Frequency GP: all energy bins GP: >1TeV July 27th, 2017 Tomás Capistrán Rojas True Energy [GeV] 5!2

24 Unfold, NNEnergyV2 + NN cut Final Unfolded Cause Distribution bin0 Frequency 8 7 NN: all energy bins NN: >1TeV 6 5 July 27th, 2017 Tomás Capistrán Rojas True Energy [GeV] 5!2

25 Flux, GPEnergyV2 + GP cut script: FluxFit.py Cause Flux Flux 11 GP: all energy bins GP: >1TeV July 27th, 2017 Tomás Capistrán Rojas True Energy [GeV] 5!25

26 Flux, NNEnergyV2 + NN cut script: FluxFit.py Cause Flux Flux 11 NN: all energy bins NN: >1TeV July 27th, 2017 Tomás Capistrán Rojas True Energy [GeV]!26 5

27 Using TWgt July 27th, 2017 Tomás Capistrán Rojas!27

28 Efficiency: NNEnergyV2 + NN cuts!28

29 Response Matrix:GPEnergyV2 + GP cut!29

30 Unfold, NNEnergyV2 + NN cut There are nan in the systematic error July 27th, 2017 Tomás Capistrán Rojas!0

Beam Test Results and ORCA validation for CMS EMU CSC front-end electronics N. Terentiev

Beam Test Results and ORCA validation for CMS EMU CSC front-end electronics US N. Terentiev Carnegie Mellon University CMS EMU Meeting, Fermilab October 21, 2005 Outline Beam Test Data vs. ORCA Simulation