Speaker
Description
The discovery of the Higgs boson in 2012 completed the Standard Model and plays a crucial role in studies of theories beyond the Standard Model (BSM). The Madala Hypothesis is one example of BSM model and puts forward a new Heavy scalar (H) to explain anomalies observed in Run 1 and Run 2 data of the LHC. The Heavy scalar H can decay to a Higgs boson in association with jets, leptons and missing energy. This scalar boson is being searched for in the $h\rightarrow \gamma\gamma + E^{miss}_{T}$ decay channel. Due to the increase in pileup interactions of the LHC, this decay channel is mostly contaminated with fake missing energy. The aim of this study is to develop a Machine Learning model to discriminate between signal events with real missing transverse energy and background events with fake missing transverse energy. We train Boosted Decision Trees (BDTs) through the Toolkit for Multivariate Analysis platform to perform this discrimination while maximising signal efficiency and background rejection. Three missing energy significance categories are considered and the impact of hyperparameter tuning is investigated. Results demonstrate that BDTs perform well and lead to a 10 - 20% improvement in accuracy.
