Identifying Dimuon Higgs Decay in Association with Top Quark Pair Using Machine Learning

Determining all measurable properties of the Higgs boson is one the ways in which we can further test the Standard Model and evaluate additional theories. Specifically, precise measurements of the Higgs coupling to second-generation fermions is the next crucial step and can be investigated via the dimuon Higgs decay channel. Here we present two methods created to examine the dimuon Higgs decay when produced in association with a top-antitop quark pair and differentiate this signal from background top-antitop pair production. The first method involves the use of an algorithm which minimizes a chi-squared-like variable. This showed some discrimination ability but was deemed insufficient for practical use. The second method employs a machine-learning algorithm and displayed correct classification of signal and background events 84 and 85 percent of the time, respectively. Future retraining the neural network with more events and a greater variety of background decays may result in accuracy rates capable of detecting signal events on ATLAS Run II and III data.