ATLAS Tau Identification in High-Multiplicity Environments
2009 (English)Report (Other academic)
The tau identification in two distinct high-multiplicity environments, pile-up and ttbar events, is investigated. The tau ID performance in such environments is known to be degraded. It is shown that it is necessary to adjust the tau ID quality cut in order to reach a working point in terms of tau efficiency versus jet rejection comparable to more frequently investigated and cleaner types of events like Z->tautau and W->taunu. Three different ways of recovering performance and robustness are investigated: Firstly, the usage of dedicated probability density functions as input to the tau likelihood method is tested, obtained from Z->tautau events with pile-up and ttbar events. Effects are generally small, except for high-pT tau jets in ttbar events. Secondly, it is investigated whether likelihood input variables can be removed without reducing performance (to increase robustness and reduce systematic uncertainties), and thirdly, whether adding new variables or replacing existing variables with new variables can lead to increased performance in terms of jet rejection at fixed tau efficiency. It is shown that several variables are redundant for tau identification in high-multiplicity environments. At least one variable is identified which can efficiently replace one of the existing variables in high-multiplicity events, namely the fraction of energy deposited in a wider isolation cone than used by the default tau identification.
Place, publisher, year, edition, pages
2009. , 31 p.
ATLAS Tau Identification
IdentifiersURN: urn:nbn:se:uu:diva-110770OAI: oai:DiVA.org:uu-110770DiVA: diva2:278220