This thesis focuses on the task of separating detector events caused by atmospheric neutrinos from those caused by atmospheric muons. Performance on this task is analysed using simulated data of...Show moreThis thesis focuses on the task of separating detector events caused by atmospheric neutrinos from those caused by atmospheric muons. Performance on this task is analysed using simulated data of these events as they are detected in the KM3NeT/ORCA10 detector setup. We present a new procedure for training the Machine Learning (ML) classifiers that handle this separation task. This most notably includes separating the data into track- and shower- like events, and training separate classifiers on these subsets of data. We show a significant improvement in the resulting neutrino signal when compared to the current classification procedure.Show less
