摘要
如何决定人工神经网络的适当规模，以往都是通过试探法来实现，不但费时，而且无规律可循。本文基于神经网络的基本学习算法，构筑动态网络结构，使之更符合抽取的新的输入、输出特性。
How to determine the appropria te size of an artificial neural netwsrk, it has usually been done through educat el guess, which either wastes time or has no rule. In this paper, based on learn ing algorithm of neural network, dynamic meural network corresponding to extract ing new features from the input and output attributes is constructed.
出处
《微机发展》
1998年第4期24-25,共2页
Microcomputer Development
关键词
动态神经网络
学习算法
神经网络
Artificial Neural Networks
Hi dden Unit
Dynamic Neural Networks
