NEURAL NETWORKS FOR HIGGS SEARCH

We describe an approach to the heavy-Higgs (m(H) = 750 GeV) search by means of a neural network (NN) in pp collisions at square-root s = 16 TeV (LHC), 40 TeV (SSC) and 200 TeV (ELN/Eloisatron). The mechanisms w e considered for Higgs production are gluon fusion and vector boson fu sion, letting the H-0 decay through the channel H-0 --> Z0 Z0 --> mumu- mu-. The overall background to the Higgs signal was assumed to con sist of the QCD continuum production of Z0 pairs, where each Z0 was fo rced to decay into muons. Using Monte Carlo simulated events at each e nergy, we trained a neural network to distinguish signal from backgrou nd and evaluated its performances as an event classifier. The results are promising and indicate that neural networks could be efficiently u sed for event selection in future experiments at super-high energy.