Performance Optimisation of tH(bb) Signal and Background Separation Using Machine Learning

The study of Higgs boson and its interaction with other particles has been one of the main topics of particles physics in the last years. Since its discovery in 2012, many experiments have been carried out in order to understand its physical properties in detail. This thesis focuses on an interaction of a Higgs boson with a single top quark, possibly the heaviest elementary particle of the Standard Model of particle physics. It uses machine-learning algorithms to filter out all unwanted processes recorded by detector to get the highest sensitivity of our target process. For this purpose, multiple machine learning models are optimized with different optimization strategies. After the best quality model is obtained, a serie of statistical tests is performed with the TRExFitter framework. The expected median value of signal strength obtained in this thesis with inclusion of statistical uncertainties was 3.89. After the inclusion of systematic uncertainties, the resulting expected median value was 6.35.