Search for the standard model production of a single top quark in association with a Z0 boson using machine learning techniques

Abstract

This thesis presents a search for the production of a single top quark in association with a Z0 boson in the dileptonic decay channel as predicted by the Standard Model. The search uses 77.8 fb−1 of data from √𝑠 = 13 TeV proton– proton collisions collected by the Compact Muon Solenoid experiment at the Large Hadron Collider during the 2016–2017 data-taking period. The search identified events containing a Z0 boson decay by requiring two opposite-sign same-flavour electrons or muons in the final state with invariant mass compatible with the nominal Z0 boson mass. Products of the top quark decay were identified using techniques developed to identify jets originating from bottom quarks and searching for a jet pair with invariant mass compatible with the W± boson mass. Machine learning techniques were used to further discriminate the signal process from background events. A study was carried out, comparing the performance of boosted decision trees with hyperparameters optimised using a Gaussian process and multi-layer perceptrons on this problem. The boosted decision trees were found to outperform the multi-layer perceptrons. A signal strength of 𝑟̂ = 6.52+2.30 −2.05 was observed, where 𝑟̂ = 1.0 corresponds to the Standard Model expectation. The corresponding observed (expected) significance is 3.12𝜎 (0.48𝜎).