神经网络法鉴别来自PP对撞不同过程的eμ事例

Neural Network Approach for Identifying Different eμEvents in pp Collision

基于ＰＹＴＨＩＡ５．６蒙特卡罗模拟，作为一种唯象的研究，对于ＬＨＣＰＰ对撞中来自Ｈ→ｒ＋ｒ－，Ｄｒｅｌｌ－Ｙａｎ，以及ｔｔ和Ｗ＋Ｗ－过程的ｅμ事例，用一具有多个输出网点的前馈式神经元网络进行鉴别，对各物理过程均获得了满意的选择效率和本底压低水平．该研究适用于未来中能质量区Ｈｉｇｇｓ粒子和ｔｏｐ夸克的寻找，以及Ｗ＋Ｗ－物理的实验研究．

Based on PYTHIA5.6 simulation and as a phenomenological study,a feed-forward neural network with multiple output neurons is built to identify and separate different ed events from H→r+r-,Drell-Yan,tt and W+ W-processes in LHC pp collision.Satisfactory selection efficiency and background suppression level are achieved.The method can thus be used in the future for experimental search for Higgs particles in intermediate energy range,top quark search,or W+W- physics study.