Speaker
Description
We will describe a study on pulse shape discrimination (PSD) with signals from Neutrino Experiment for Oscillation at Short baseline II (NEOS II). In an experiment using a liquid scintillator detector, such as the NEOS II, PSD plays an important role in distinguishing the signals of electron-induced events from those of fast neutron-induced events to improve the signal-to-background ratio. The conventional PSD method uses the ratio of the tail part to the total area of the waveform. Instead, we applied a convolutional neural network (CNN) trained from Fast Fourier-Transformed waveforms of alpha and beta particles selected in the NEOS II data. The trained CNN model was evaluated with the Na gamma source data and showed an improvement in the signal-to-background ratio compared to the conventional method results.
| Collaboration | NEOS II |
|---|
