一种基于概率神经网络模型的胃病诊断专家系统  

A Neural Network Model Based on Bayesian Posterior Probability

作　　者：吴新根 柏毅 

机构地区：东南大学生物科学与医学工程系

出　　处：《东南大学学报（自然科学版）》1999年第4期57-62,共6页Journal of Southeast University：Natural Science Edition

摘　　要：给出了将神经网络与基于规则的专家系统结合起来的一种神经网络模型，在这种模型中，一方面，领域知识以显式方式表示，神经网络的并行推理过程是可以理解和解释的；另一方面，由于通过贝中斯后验概率指定权值，因此避免了神经网络漫长的递归学习过程，提高了学习效率。In this paper we propose a neural network paradigm that combines neural networks and rule based expert systems, in which domain knowledge is encoded explicitly in the form of rules, and the parallel reasoning processes can be understood and explained. On the other hand, due to the using of Bayesian posterior probability to determine the connection weights, we avoid the long training process of neural network and improve the learning efficiency. The empirical test shows the model can not only express domain knowledge, but also carry out uncertainty parallel reasoning.

关 键 词：神经网络 专家系统 胃病诊断 不确定性推理 

分 类 号：R573.04[医药卫生—消化系统] TP18[医药卫生—内科学]