摘要
本文提出一种训练标准ＢＰ网络的方法，该方法能以模糊判定规则的形成立即解释连接权值。该方法可以用来从一组例子中生成全新的分类规则。
This paper presented a method for training standard BP networks in such a way that the learnt conection weights can immediately be interpreted in fuzzy decision rules.This can be applied to refine an existing rule base as well as to induce totally new classification rules from a set of examples.
出处
《小型微型计算机系统》
EI
CSCD
北大核心
1997年第6期76-78,共3页
Journal of Chinese Computer Systems
关键词
BP算法
神经网络
符号规则
BP algorithm, Neural networks, Symbolic rules
