Classification of hadronic events in e+e− collisions with neural networks: applications to tag the quark and the number of jets

Abstract The classification by a multilayer perceptron (MLP) of e + e − events according different classes is presented. The introduction of variables connected with the vertex detector improves the tagging of the quark of the event. The tagging of the number of jets is also considered. New 3-D invariants are presented in the appendix.