Neural network approach to obtain the primary proton flux at the knee from a hybrid-experiment of EC and AS array in Tibet

A hybrid experiment of emulsion chamber and air shower array will be started from 1996 at Yangbajing (4300 m above sea level) in Tibet, aimed at obtaining the primary proton flux around the knee energy region. In order to examine a classification power of proton induced showers from others using this apparatus, we have done an extensive Monte Carlo simulation in which the observation conditions were adequately taken into account. For this purpose, a multiple layered feed forward neural network was used as a classifier to select proton and helium induced shower events preferentially. It is shown that the separation of protons and the proton-helium group from all events can be achieved with an efficiency of 80–85%. The performance of the experiment under a two-year exposure is discussed.