Summer Student Report 2015. Title : Separating prompt $B_s^0$ from secondary $B_s^0$ originating from a $B_c^+$ using machine learning. Author : Blaise Delaney, Trinity College, University of Dublin. Supervisor: Dr Matthew Kenzie, CERN.

This report describes the project carried out under the supervision of Matthew Kenzie in association with the LHCb (Large Hadron Collider Beauty Experiment) collaboration at CERN. The project entailed developing a machine learning (ML) algorithm capable of differentiating between the prompt $B_s^0$ production and the secondary $B_s^0$ production originating from a $B_c^+$, in order to estimate the production fraction, $\frac{f_c}{f_s}$. By carrying out our analysis on Monte Carlo simulated decays sharing the same final state $J / \psi K^+ K^-$ it was possible to separate the $B_c^+$ production from the prompt $B_s^0$ production with low systematic uncertainties, attaining a final ROC score of 0.6957.