一种基于神经网络逆向推理机制的专家系统

本文提出了一种基于神经网络逆向推理机制的专家系统设计方案。论文研究了一种表达知识的二元产生式规则及其编码方法。通过编码，知识被存储在ANN中；基于ANN的知识库，论文设计了一种具有逆向推理机制的推理机。本文设计了一个原型系统，用于动物识别，经实验，取得较理想的效果，证明了该方法的有效性。

The Design of Expert System Based on Neural Network Backward Inference

This paper proposes a design of expert system based on neural network backward inference . The knowledge expression adopts production rules. Knowledge is not only standardized into binary rules, but coded to store in the neural network. According to the binary rules, the paper proposes a backward inference algorithm based on neural networks. We design a prototype system used in animal recognition and get a good result, which proves its effectiveness.