Wing Sheung Chan

Object reconstruction and identification 65 Table 3.2.: Physics object selection criteria that are specific to the search for LFV Z → `τ decays. Object Selection criteria Jet p T > 20 GeV | η | < 2 . 5 b -jet Passes fixed cut on the MV2c10 output corresponding to a 77% b -tagging efficiency τ had - vis p T > 25 GeV | η | < 1 . 37 or 1 . 52 < | η | < 2 . 5 Tight RNN ID Medium ( Loose ) e -veto for 1-prong (3-prong) candidates Electron p T > 30 GeV | η | < 1 . 37 or 1 . 52 < | η | < 2 . 47 Tight identification FCTight isolation Muon p T > 30 GeV | η | < 2 . 5 Tight identification FCTight isolation E miss T Tight working point Track-based soft term The reconstructions of jets, electrons, muons and τ had - vis are carried out independently without much harmonisation. Consequently, the same tracks and energy deposits may be attributed to multiple objects, leading to inflated object multiplicities, and incorrectly reconstructed event topologies and kinematics. To resolve this, objects that share the same ID track with or are geometrically close to another object are removed in a specific, well-defined order. This conflict-removing procedure is called overlap removal. For the purpose of overlap removal, a looser definition for electrons, muons and τ had - vis is used, which helps to reduce the misidentification rates. For electrons and muons, the p T threshold is lowered to 25 GeV , the identification requirement is relaxed to Loose and no isolation requirement is enforced. For τ had - vis , the RNN ID requirement is relaxed to Loose and no electron rejection is applied.