Machine Learning Approaches for Background Modeling and Dimuon Classification in the SpinQuest Experiment

The E1039/SpinQuest experiment at Fermi National Accelerator Laboratory uses a 120~GeV proton beam from the Main Injector, incident upon transversely polarized protons and deuterons using ${NH}_3$ and $ND_3$ targets, respectively. In addition to measuring the Sivers asymmetry in Drell--Yan ( pp ) and ( pd ) scattering from sea quarks, SpinQuest will study transverse-spin effects, particularly the transverse single-spin asymmetry (TSSA) in $J/\psi$ production. The angular distributions from the $J/\psi$ decay could play an important role in understanding the gluon contribution to the proton spin structure. To extract asymmetry, it is critical to isolate dimuons from specific physics channels and separate combinitoric background. For accurate classification of signal events using a machine learning-based classifier, a large sample of simulated background events is essential for training. We ensure that background distributions are closely tuned and matched with the experimental background from commissioning data to enable reliable classification. Ultimately, signal events, such as $J/psi$ will be used to extract the TSSA.