摘要
用神经网络为非线性软测量建模 ,用于推断估计不可在线测量的变量。利用训练样本集的单调投影指数直接估算隐层神经元数目 ,有较为满意的逼近精度。
Modeling for nonlinear soft measuring by means of neural network can be used to estimate variables that can not be measured on line.A monotone index based methods is presented to directly estimate the number of hidden neurous in feedforward neural networks.This method can smoothly approximate the training data and provide a satisfactory degree after trainning.
出处
《上海工程技术大学学报》
CAS
2000年第2期88-92,共5页
Journal of Shanghai University of Engineering Science
