Testing lepton flavour universality in $B\rightarrow K\pi\pi\ell\ell$ decays using 2017 LHCb data

This present work proposes a study on $B\rightarrow K\pi\pi\ell\ell$ decays using 2017 LHCb data in the context of Lepton Flavour Universality tests. In view of the determination of the ratio of branching fractions $R_{K\pi\pi}$, that complements the actual investigations around the Standard Model shortcomings, several intermediate computations have been carried out. A previous work provided two corrections, concerning the particle identification, and the kinematics and multiplicity variables, associated to the Monte Carlo simulation. These are useful in the accurate separation of events from a background of incorrectly reconstructed candidates; a task accomplished by using the boosted decision tree (BDT), a multivariate analysis (MVA) method, trained and tested on specific data sets defined for a selected set of discriminating features. The resulting classifier output, providing the classification, has an optimised threshold determined by a figure of merit that yields a signal efficiency of up to 82.1\% while removing more than 92.5\% of background.