Electron Identification with the ATLAS detector

Electron identification for $pp$ collisions at $\sqrt{s}=$ 13 TeV with the ATLAS detector is performed using a likelihood based method both online in the trigger and offline. The likelihood is an improvement on the cut based method. A multivariate analysis technique (MVA), the likelihood method is able to simultaneously assess many characteristics before making a decision. This likelihood is constructed from variables that discriminate between electrons and backgrounds. As some of these variables depend on the number of pileup collisions per bunch crossing, it is advantageous to have an event-by-event measure of the number of pileup collisions. This can be used to adjust the likelihood discriminant value as a function of the pileup to ensure that the electron identification remains efficient at high pileup, without drastically increasing the amount of background accepted. This is currently being done with the number of primary vertices offline and the average number of collisions online as the measures for pileup. A potential improvement presented here is to measure the pileup with the TRT local Track Occupancy, which is a more local descriptor of activity around an electron candidate and is available both online and offline.