Casasent David, Chen Xue-wen
Department of Electrical and Computer Engineering, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
Neural Netw. 2003 Jun-Jul;16(5-6):529-35. doi: 10.1016/S0893-6080(03)00086-8.
We propose a novel technique for the design of radial basis function (RBF) neural networks (NNs). To select various RBF parameters, the class membership information of training samples is utilized to produce new cluster classes. This allows emphasis of classification performance for certain class data rather than best overall classification. This allows us to control performance as desired and to approximate Neyman-Pearson classification. We also show that by properly choosing the desired output neuron levels, then the RBF hidden to output layer performs Fisher discrimination analysis, and that the full system performs a nonlinear Fisher analysis. Data on an agricultural product inspection problem and on synthetic data confirm the effectiveness of these methods.
我们提出了一种用于设计径向基函数（RBF）神经网络（NN）的新技术。为了选择各种RBF参数，利用训练样本的类别隶属信息来生成新的聚类类别。这使得能够强调某些类别数据的分类性能，而不是最佳的整体分类。这使我们能够根据需要控制性能并近似奈曼-皮尔逊分类。我们还表明，通过适当选择所需的输出神经元水平，RBF隐藏层到输出层执行费舍尔判别分析，并且整个系统执行非线性费舍尔分析。关于农产品检测问题和合成数据的数据证实了这些方法的有效性。
