关键词: 神经网络 专家系统 知识表示 知识获取 推理

Abstract:

In allusion to the insufficiencies such as the weak inference capability and low intelligence level of traditional expert systems, this paper solves t he questions of knowledge representation and knowledge acquisition of traditional expert systems based on neural networks. The paper begins with the several insufficiencies of traditional expert systems, expatiates on the fundamental theory and the framework of expert systems based on neural networks, and finally gives the implementation of a fault-diagnosis system for artesian wells by selecting a three-layer BP neural network model.

Key words: neural network;expert system;knowledge representation;knowledge acquisition;inference