Revisiting new and old jet clustering algorithms for beyond the standard model Higgs searches in the final states with b-jets

The search for novel physics Beyond the Standard Model (BSM) continues to be elusive despite the Large Hadron Collider’s (LHC) many triumphs since its inception in 2008. The ultimate aim of this work is to address this issue and search for new physics using the simplest extended Higgs sector framework, the 2- Higgs Doublet Model (2HDM), manifested in cascade decays with high multiplicity b-jet final states wherever kinematically possible. In this thesis, we compare different jet clustering algorithms to fully resolve hadronic b-jet final states arising from a decay chain of a heavy CP-even Higgs H into a pair of the lighter Higgs bosons h. We consider both scenarios where mH > mh = 125 GeV and mH = 125 GeV > mh for the 2HDM Type-II framework. We provide the ideal choice of acceptance cuts, resolution parameters and reconstruction procedures in order to enhance the significance ratios and establish such a ubiquitous BSM signal using the 2HDM Type-II framework. Furthermore, we examine the potential of detecting a cross-section at the High-Luminosity phase of the LHC (HL-LHC) for the production of SM-like h in asssociation with a single top quark. For the illustrative example of bg → twh with h → b ¯b final state, the permissible benchmark points in the 2HDM Type-II are shown to yield better significance rates and distinct kinematical distributions with respect to the SM, allowing the signal to be observed at the HL-LHC.Finally, we employ the machine learning method of image recognition to design a Convolutional Neural Network (CNN) to classify the double b-tagged fatjet final states emerging from a 2HDM Type-II signal against the leading background