Selecting Electron Neutrino Events in the MINOS Detectors

The MINOS Collaboration recently completed a search for $\nu_e$ appearance in the NuMI beam at Fermilab. Since obtaining the first result, we have worked on improving the particle identification algorithms that distinguish $\nu_e$ charged current events from various background events. These include ANN (Artificial Neural Network) and LEM (Library Event Matching). ANN is a neural network that uses a set of reconstructed quantities that characterize the longitudinal and transverse energy deposition profiles of a given event. LEM is a pattern-recognition algorithm that compares the hit pattern of a given event to the hit patterns of many simulated ``library'' events; it then constructs discriminant variables from those library events that best match that event. The development of particle identification algorithms of such fundamentally different natures allows us to make a cross-check of our results. Event topologies in the detectors are discussed, and the effectiveness of these algorithms in selecting electron neutrino events is examined.