Quark- and gluon-jet separation using neural networks

We present a possible discrimination method by the combination of a neural network (NN) and QCD to separate the quark and gluon jets of e+e- annihilation. The network has been trained by jets of the same energy; thus only the internal structure of the jets has been considered. By the combination of the NN and the QCD matrix element, 92% accuracy in identifying jets has been achieved.