The DNNCascades event selection is a neural network based, high statistics cascade-like dataset that was first used to detect high energy neutrinos in the Galactic plane. As well as contained cascades, the selection includes ~30% uncontained cascades – neutrinos with interaction vertices at the edge or outside of the detector instrumentation volume. The high statistics, contained and uncontained cascades event selection could be key to more tightly constraining the diffuse flux across the energy spectrum.
My work involves optimizing this unique event selection to bring into the diffuse neutrino physics space. Extensive work on updated atmospheric neutrino background modeling, systematics updates, and data/MC improvement will be discussed, as well as my intention to perform an astrophysical diffuse flux measurement that could resolve uncertain features of the astrophysical spectrum.