guest ::
| Home > Publications database > Studies of top tagging identification methods and development of a new heavy object tagger |
| Home > Publications database > Studies of top tagging identification methods and development of a new heavy object tagger |
| Report/Dissertation / PhD Thesis | PUBDB-2016-02025 |
; ;
2016
Verlag Deutsches Elektronen-Synchrotron
Hamburg
This record in other databases:
Please use a persistent id in citations: doi:10.3204/DESY-THESIS-2016-013
Report No.: DESY-THESIS-2016-013
Abstract: At the Large Hadron Collider (LHC), precision tests of the standard model of particle physics and searches for new phenomena are performed. To make optimal use of the proton-proton collisions delivered by the LHC and its increasing collision rate, both the detectors and the reconstruction algorithms have to be optimized. The identification of heavy quarks is a key component in many measurements. This thesis describes a hardware and a software project which both aim at improving the identification of heavy quarks.In the first part of this thesis, the Phase 1 upgrade of the CMS pixel detector is introduced. One of the main motivations of the replacement of the Pixel detector is the improved b jet identification at large collision rates. The Phase 1 upgradeinvolves several production and calibration steps. An x-ray calibration procedure has been developed and the corresponding experimentalsetup is described. Measurements show that the calibration of the pixel modules is temperature independent and can be performed at room temperature. The stability of the setup is tested in order to fulfill the requirements for mass production of the pixel modules. A method to stabilize the calibration is introduced which is shown to reduce the systematic uncertainty.In the second part, algorithms to identify heavily boosted top quarks (``top tagger'') are described and their performance is compared. The OptimalR HEP top taggerand the shower deconstruction tagger show a better performance than existing tagging algorithms. They can be used in Run II with increased centre-of-mass energies of 13 and 14 TeV. It is also shown that existing top tagging algorithms can be improved by the usage of multivariate analysis methods. New algorithms are commissioned using CMS data with a centre-of-mass energy of 8 TeV, corresponding to an integrated luminosity of 19.7 $\rm{fb}^{-1}$.In order to validate these new algorithms in data,two selections are made to measure the efficiency and mistag rate. The selections aim at enriching one sample with $\rm{t\overline{t}}$ events and one with QCD multijet events. In the last part of this thesis, a novel top tagging algorithm is introduced. It is capable of identifying top quarks over a large $p_{\rm{T}}$ range by using a $p_{\rm{T}}$ dependent jet size and a method to reconstruct subjets based on a mass jump criterion. The algorithm includes the possibility to remove soft and collinear radiation from the jet. The algorithm is infrared and collinear safe, has a reasonable computational time, and shows a good performance with respect to existing algorithms. The performance of the new algorithmis also investigated on 8 TeV data.
|
The record appears in these collections: |
PDF
PDF (PDFA)