本文论述了一个以激活与竞争神经网络为基础的多专家意见综合系统,它是一个可存储规则并可学习的系统,常规的加权法相当于神经网不存储任何知识时的特例,该系统在听取了各位专家意见后,即进入一个自我封闭的静思过程,然后做出判断,这种处理多专家意见的方式与人脑非常相似,很值得研究。 This paper discusses a comprehensive multi-expert opinion system based on activation and competition neural networks, which is a system that can store rules and can learn. The conventional weighting method is equivalent to the special case when the neural network does not store any knowledge. After listening to the opinions of experts, we enter a self-enclosed meditation process and make judgments. This way of handling multi-expert opinions is very similar to that of the human brain and is worth studying.