摘要
智能系统中一类重要的定性知识要用模糊集理论中的模糊语言进行描述．本文在研究模糊定性知识形式描述和自组织竞争神经网络特性的基础上，提出了一种从一组具有数值特性的训练样本集中获取隶属函数和模糊规则的神经网络模型和方法．通过对Ｉｒｉｓ数据集的应用实验表明了该方法能对这一类数据进行有效的描述．
In the intelligence system,a kind of important qualitative knowledge is described by the fuzzy language of the fuzzy set theory. On the basis of the fuzzy qualitative knowledge formal description and self-organization competitive neural network characteristic, this paper presents a neural network model and method to acquire the membership functions and fuzzy rules from a training sample set of the numerical features. The experimental results about the Iris data set show that this method can describe these data effectively.
出处
《信息与控制》
CSCD
北大核心
1998年第6期408-412,共5页
Information and Control
基金
辽宁省自然科学基金
