摘要
粗集和神经网络都是处理不确定、不完全信息的软计算方法 ,两者曾在决策支持和知识获取等领域取得了很大的成功并得到初步应用。但两者都有局限性 ,同时在许多方面有互补性。因此粗集和神经网络的集成成为当今智能混合系统的一个重要分支 ,或许也是开发下一代专家系统的主流技术。从认识论的角度分析了粗集和神经网络的特点 ,评述了目前粗集和神经网络集成的理论和方法 ,指出其中存在的主要不足。最后提出了粗集和神经网络进一步集成所需解决的问题。
Rough set and neural networks are both soft computing methods dealing with indefiniteness and incompleteness. They once played an important role in decision support, knowledge acquisition and so on. But there exist some limits when they are simply applied. On the other hand, because they are complementary, the integration of rough set and neural networks has become a hot topic in the domain of intelligent hybrid systems with great practical values and broad developing prospects for the designing of expert systems. The characteristics of rough set and neural networks are firstly analyzed in terms of epistemology and an overview of their integration and some comments are made in this paper. Furthermore, some problems concerning further research are presented.\;
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2002年第10期103-107,126,共6页
Systems Engineering and Electronics