Download presentation
Presentation is loading. Please wait.
1
1 CC analysis update Status of the cross-section reweighting package Status of the Physics Analysis Ntuple (PAN) D. A. Petyt Nov 3 rd 2004
2
2 Cross-section reweighting package A working version of this package now exists (thanks to Chris Smith and Hugh Gallagher) –The package provides an interface to NEUGEN and allows the user to calculate a weight for a particular GMINOS event given a tweaked set of NEUGEN input parameters. –The parameters are: M A for quasi-elastic scattering M A for resonance production DIS/Resonance scaling factors (8 parameters – 2 multiplicities, 4 initial states) Parton distribution functions (from PDFLIB) –This package will enable us to assess the effect of cross-section uncertainties on the CC analysis It is also required for the completion of the Mock Data Challenge –The code is still under development, but should be available for general use soon. The source code currently lives in the Contrib area of the minossoft CVS repository under Contrib/cbs/NeugenInterface
3
3 Cross-section reweighting function The weighting function has the following form: Double_t reweight( Float_t E_nu, Kinematic_variable_t::kv1, Float_t val1, Kinematic_variable_t::kv2, Float_t val2, Interaction I, process_t::proc neugen_config newconfig neugen_config oldconfig ) Neutrino energy Kinematic variables (x,y,q 2 or W) relevant to this interaction and their values Includes: flavour, nucleus (A,Z), CC or NC, initial state ( p, n, N …) QEL, RES, DIS, COH These objects hold the values of the neugen config parameters (m A etc). The event weight is calculated based on these two sets of quantities
4
4 The Physics Analysis Ntuple (PAN) A simple ROOT ntuple that contains “all relevant quantities” for the CC analysis. –Truth information for MC events, including sufficient information to permit event reweighting (i.e. need at minimum E_nu, x, y, q 2, W, E_mu, limited STDHEP information). Can be extended to accommodate other reweighting functions as required. –Reconstructed information necessary to perform oscillation fit –Quantities specific to the CC analysis, such as PID parameters and fiducial cuts, that enable CC-like events to be selected –Association of tracks/showers/slices in the Near Detector. Identical format for ND and FD events An initial version of the PAN now exists with the quantities necessary for cross-section reweighting. A sample PAN for the far detector can be downloaded from the following website: –http://www.physics.umn.edu/~petyt/cc/PAN/ Ed Lartey has provided code to produce a PAN object from within the Mad analysis framework
5
5 PAN variables - 1 Variable Description true_enutrue neutrino energy (GeV) true_pmu true muon momentum (GeV/c) true_ehad true hadron energy (GeV) true_mudircos true muon z-direction cosine cc_nc cc/nc flag: 1-cc 2-nc flavour true flavour: 1-e 2-mu 3-tau process process: 1001-QEL 1002-RES 1003-DIS 1004-COH true_x true x true_ytrue y true_q2true q^2 true_wtrue w initial_state initial state: 1-vp 2-vn 3-vbarp 4-vbarn... nucleus target nucleus: 274-C 284-O 372-Fe run_num run number snarl snarl number evt_index event index within snarl mc_index mc index for this event numevt number of events/snarl numtrk number of tracks numshw number of showers evtvtx event vertex x-coord (m) Required for cross-section reweighting
6
6 PAN variables - 2 evtvty event vertex y-coord (m) evtvtz event vertex z-coord (m) is_fid vertex fiducial volume flag: 0-fail 1-pass is_cev containment flag: 1-CEV 2-PCE trkpass track quality flag: 0-fail 1-pass evlength event length (planes) trklength track length (planes) trkrange track momentum range (GeV/c) trkqp track momentum q/p trkeqp error on fitted q/p trkdircos track z-direction cosine shwph summed shower pulse height (shw.ph.gev) reco_pmu reconstructed muon momentum (GeV/c) (range or q/p) reco_eshw reconstructed shower energy (shw.ph.gev/1.23) reco_enu reconstructed neutrino energy (GeV) phfrac track pulse height fraction (uses sigcor) phplane track pulse height per plane ( " " ) pid1 pid parameter - Super-K definition pid2 pid parameter (probmu/(probmu+probnc)) neural neural network output is_cc cc-like flag: 0-nc-like 1-cc-like The only cut required to select cc-like events
7
7 Reweighting FarDet events – a simple example QEL: change ma_qel from 1.01 to 1.1 GeV RES: change ma_res from 0.95 to 1.05 GeV DIS: multiply DIS/RES matching scale factors by 0.75
8
8 PID variable by process All events QEL NC DIS RES is_cc=1 cut applied
9
9 Effect of weighting Pid parameter unweighted weighted
10
10 Near/far comparison – change m A by 10% nominal m A -10% m A +10%
11
11 Far/near ratio Effect largely cancels out in far/near ratio
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.