Evidence for the associated production of the Higgs boson and a top quark pair with the ATLAS detector

This thesis presents analyses performed with proton-proton collision data collected with the ATLAS detector in $2015$ and $2016$ at $\sqrt{s}=13$ TeV. The measurement of so-called non-factorisation effects in the calibration of luminosity is presented. An estimate of the correction, and systematic uncertainty, due to non-factorisation on the calibration is calculated by measuring the proton bunch density profiles from the variation in the luminosity and distribution of reconstructed vertices during beam separation scans. The correction is applied to the calibrations of the total luminosity collected in both 2015 and 2016. A novel multivariate algorithm designed to reject non-prompt leptons (produced from the decays of $b$- and $c$-quarks) is presented, utilising information from nearby tracks to discriminate from prompt leptons (produced from $W$, $Z$ and $H$ boson decays). This algorithm is used to reject non-prompt backgrounds in the search for the associated production of a top quark pair and a Higgs boson ($t\bar{t}H$) in multilepton final states with $36.1$ fb$^{-1}$ of $\sqrt{s}=13$ TeV data. Multilepton states refer to the Higgs boson decaying into pairs of $W$ bosons, $Z$ bosons or $\tau$ leptons. The combination of the multilepton analysis with the other search analyses of $t\bar{t}H$ production in which the Higgs decays to pairs of photons, $b$-quarks and $ZZ\rightarrow4\ell$ is also shown. The measured value of the signal strength of $t\bar{t}H$ production in data is $\mu(t\bar{t}H) = 1.2 \pm 0.3$, corresponding to an observed (expected) discovery significance of 4.2$\sigma$ (3.8$\sigma$) and constituting evidence for the $t\bar{t}H$ production mode.