【摘要】 论文以北京正负电子对撞机上BES实验中的e、μ、π粒子的神经元网络法鉴别为例 ,通过对原始输入变量进行PCA(PrincipalComponentAnalysis)方法重组 ,构造出一组相互独立的输入变量 ,去除原始变量中的某些相关性 ,可以减少输入变量的个数 ,改善网络的操作速度性能 .更多还原

【Abstract】 In the identification of e, μ and π particles at Beijing Spectrometer (BES) experiment with neural networks, the original input variables are reorganized by using PCA (Principal Component Analysis) method to form a set of irrelevant ones. The correlations between original input variables are thus eliminated, and the number of input variables is reduced. A higher speed and better operation capability are achieved.更多还原