确定人工神经网络中间层神经元数的方法—奇异值分解

SINGUUR VALUE DECOMPOSITION AS A METHOD FOR DETERMINATION OF HIDDEN NUMBER IN ARTIFICIAL NEURAL NETWORK

人工神经网络（ＡＮＮ）在化工中的应用日益广泛，这些应用包括过程控制系统的故障诊断、非线性动态的模拟、非线性系统的控制和非线性规划问题的求解等。大部分发表的文献集中在ＡＮＮ的应用上。作者在介绍ＡＮＮ的基本原理后，主要探讨ＡＮＮ的性质，提出了将矩阵奇异值分解作为确定ＡＮＮ中间层节点数目的方法，对中间层节点数目与所处理问题变量之间的可能关系作了讨论。

The artificial neural.network (ANN) has been used for important applications in chemical engineering. Such applications include: fault detection in process control systems. modeling ofnonlinear dynamics, control of nonlinear systems,. solution of nonlinear programming problems andothers. Most of the articles published on the subject concentrate on the applications. After introducing the bssic principles of ANN, we concentrate on the properties of ANN. In particular, we studythe method for the determination of hidden number in ANN by using Singular value decomposition(SVD). With this method, the possible relstion between the hidden number of ANN and the variables of the problem is discussed.