How are cone jets reconstructed ? - PowerPoint PPT Presentation

1 / 6
About This Presentation
Title:

How are cone jets reconstructed ?

Description:

Needed to create a class (ReadNoiseFile) to read in the noise file ... Apparently number of fake jets with 2.5 sigma ZS. small ... – PowerPoint PPT presentation

Number of Views:12
Avg rating:3.0/5.0
Slides: 7
Provided by: wwwd
Category:
Tags: cone | create | jets | reconstructed | run | up

less

Transcript and Presenter's Notes

Title: How are cone jets reconstructed ?


1
Emmanuel Busato
How are cone jets reconstructed ?
Vishnu's proposal to improve reconstruction
At preclustering level not let a CH cell
become a seed if it's
energy is
less than some sigma
2
Current preclustering (Simple Cone)
3
Modified preclustering
4
Implementation
Everything is on coco-clued0 /work/coco-clued0/b
usato/jetanalyze_p13.06.01 - Major modifications
are in SconeClusterAlgo.hpp - Needed to create a
class (ReadNoiseFile) to read in the noise file
(energycluster/energycluster/ReadNoiseFile.hpp)
ReadNoiseFile uses official cal_nlc and
em_util packages methods to . convert channel
(0 ??55295) ? (ieta, iphi, layer) .
convert ADC ? GeV (problem with GeV ? rawADC)
Current algo works with linearized ADCs gain 8,
should we work with raw ADCs and/or gain 1
? - 3 new parameters in the rcps.

5
CalPreSCILCone05(07,03).rcp
string PackageName "CalClusterReco" string
data_type "MC" string algo_type
"PreSCilcone" bool drop_negative_mass2
true // preclustering parameters float coneSize
0.3 float seedET 0.5 float pTmin
1.0 int minItems 2 bool dropCHseed
false float dropCHseedsigma 3.5 string
offline_zero_supp_file "/work/coco-
clued0/busato/jetanalyze_p13.06.01/cal_zsup_noise.
dat" // cone clustering parameters float
Item_ET_threshold 0. float Radius_of_Cone
0.5 float Min_Jet_ET 8.0 float ET_Split_Frac
0.5 float Far_ratio 0.5 float ET_min_ratio
0.5 bool Kill_duplicate true float
Duplicate_dR 0.005 float Duplicate_dPT
0.01 // if useMCvertex is true use Monte Carlo
vertex // if useD0vertex is true use
reconstructed vertex // if both true use Monte
Carlo vertex if there is no reconstructed
vertex bool useMCvertex false bool
useD0vertex true
To run the algo with modified preclustering,
set dropCHseed to true
6
Status
- This modification was first made to reduce the
number of fake jets, but - good jets can have
high f90 - Apparently number of fake jets with
2.5 sigma ZS small More to come
to confirm this (jets in dijet events study). -
Could be used instead to reduce the effect of
noise on good jets
MC study done by Vishnu
Write a Comment
User Comments (0)
About PowerShow.com