Ponente
Descripción
The IceCube Neutrino Observatory is a cubic kilometer-sized detector designed to detect neutrinos of astrophysical origin. We discuss a new method using a boosted decision tree to classify and reduce the cosmic ray muon rates from ~billion per year to ~1 per year, the same BDT is also used to select for starting track events (~900 per year). The classification and removal of incoming muons is also efficient at removing atmospheric neutrinos with accompanying muons allowing us to characterize the astrophysical neutrino flux at energies between 1-100 TeV in the southern sky. The high purity starting track dataset has two advantages: excellent energy and angular resolution. The presence of the hadron and muon allow us to directly reconstruct the energy of the neutrino due to the vertex being well contained within the detector volume. A Random Forest is used resulting in a neutrino energy resolution below 30%. The outgoing muon track is then used to reconstruct the neutrino direction with an estimated median angular resolution of 1.4° at 1 TeV improving to 0.7° at 100 TeV. These advantages are particularly useful in this high-statistic starting track dataset where detector and theoretical systematics begin to dominate over statistical uncertainty.
