Bonn 2020 – wissenschaftliches Programm
Die DPG-Frühjahrstagung in Bonn musste abgesagt werden! Lesen Sie mehr ...
T 70.4: Vortrag
Donnerstag, 2. April 2020, 17:15–17:30, H-HS I
Standard Model H→ττ analysis with a neural network trained on a mix of simulation and data samples — Günter Quast, •Moritz Scham, and Roger Wolf — Karlsruhe Institut für Technologie, Karslruhe, Deutschland
At the LHC, best access to the coupling of the observed Higgs boson to fermions is provided by the decay into τ leptons. For the measurement of simplified template cross sections in this decay channel CMS uses a multi-classification neural network to distinguish signal from background classes. In this talk a training of the neural network is presented, where the most important backgrounds from genuine di-τ events and from jets misidentified as τ leptons are obtained from data. In this way up to 90% of the background events are estimated from data, independent from the simulation.