Peking Univ, Dept Tech Phys, Sch Phys, Beijing 100871, Peoples R China
; CAS, Inst High Energy Phys, Beijing 100039, Peoples R China
The final states created by e(+)e(-) annihilation at tau-charm energy region are simulated with PYTHIA5.7, and the resultant inclusive muon production cross sections and momentum distributions are obtained. The muon identifier of BES M detector is fully simulated with Geant3.21. For the simulated events containing muons and pions with momentum greater than 0.5GeV/c, the BP neural network and Fisher discriminant method are used for muon/pion identification. BP neural network performs better than Fisher discriminant method. With the same muon efficiency of 96%, the misidentification of pion by BP neural network is 9.4%, in comparison with that of 17.1% by Fisher method.