论文部分内容阅读
提出了用规则学习方法得到的规则建立人工神经网络的方法基于规则的神经网络方法。它是一种规则学习方法和人工神经网络各自取长补短有机结合的方法。它即避免了规则学习方法的识别速度慢和神经网络训练慢的缺点，又保留了规则学习方法的训练速度快、聚类能力强和神经网络的识别速度快、可硬件实现的优点。它为大规模训练样本来建立神经网络识别系统开辟了一条新的途径。
The method of rule-based learning is proposed to establish the artificial neural network method based on the rule-based neural network method. It is a method of learning by law and artificial neural network to learn from each other. That is to say, it avoids the disadvantage of regular learning method recognition and slow training of neural network, and retains the advantages of rule-based learning methods such as fast training speed, strong clustering ability and neural network recognition speed and hardware implementation. It opens up a new way for large-scale training samples to establish neural network identification system.