Electron/pion identification with ALICE TRD prototypes using a neural network algorithm

We study the electron/pion identification performance of the ALICE Transition Radiation Detector (TRD) prototypes using a neural network (NN) algorithm. Measurements were carried out for particle momenta from 2 to 6 GeV/c. An improvement in pion rejection by about a factor of 3 is obtained with NN compared to standard likelihood methods.