Possibility of neural network usage for $B$-meson events selection in calorimetric trigger

With Monte Carlo generated model data a possibility of feed-forward neural network usage for $B$-meson events selection in calorimetric trigger of multyparticle spectrometer MPS was investigated. A comparison of indicated events selection with the help of neural network and traditional method based on distinctions in distributions of transfer energy have shown an advantage and perspectivity of a new approach.