Boosted Higgs boson tagging using jet substructures

Searching BSM particles via the Higgs boson final state has now become common. The mass of desired BSM particle is more than 1 TeV, thereby its decay products are highly Lorentz-boosted. Hence the jets from b quark-antiquark pair - which the Higgs boson mostly decays into - are very closed to each other, and merged into one jet, that is typically reconstructed using large jet sizes (∆R = 0.8). In this work regression technique is applied to AK8 jets (which defined by anti-kT algorithm, using ΔR = 0.8). The regression makes use of boosted jets with substructure information, coupled with the pecularities of a b quark decay, like the presence of a soft lepton (SL) inside the jet. It has allowed to improve the resolution of the mass reconstruction and transverse momentum of the Higgs boson. This application results in improvement of the mass reconstruction by 3-4 percent. These result may be improved firstly by making more careful pileup rejection. Then it is possible to combine base regression train for different BSM particle mass. Finally other decays may be considered for example one with diboson or H + t production.