一种基于符号神经网络的知识获取方法

本文介绍一种基于符号神经网络的知识获取方法，该方法首先用传统的机器学习方法获取关于某领域的粗略知识，然后把这些知识映射到神经网络结构，通过神经网络的自学习获取关于该领域的精细知识，这样，既解决了传统机器学习中知识精度，知识表示等问题，又解决了神经网络获取知识时间长，能释能力弱等问题。

A METHOD OF KNOWLEDGE ACQUISITION BASED ON SYMBOLIC NEURAL NETWORKS

In this paper, a methofd of knowledge acquisition based on symbolic neural networks is introduced, First the coarse knowledge of a field is acquired with the customary machine learning method. then the knowledge is mapped to neural network structure, at last the fine knowledge of a field is acquired through the auto learning of neural network in this method. Thus. the problems of knowledge refinement in machine learning method and the problems of long knowledge acquisition time and weak interpretation ability in neural networks are solved.