摘要
<正>我们知道,在一般的代数空间中,集合X和Y的二元关系定义为XxY的子集.2~Inan在研究人的视觉结构中,从一般拓扑空间理论出发建立了容限空间理论。在容限空间理论中,定义了容限、容限空间等概念。这些概念的提出对于我们正确认识人工神经网络功能有着重要的指导意义。
TST (Tolerance Spaces Theory) suggested by Zmoeeman has defined some concepts such as Tolerance and Tolerance Space. Based on known knowledge on TST, this paper mainly discusses the following problems; 1) the further study of TST, specially in the definition of Generations; 2)TST is applied to some problems of Artificia Neural Networks; 3) an algorithm which can be used for machine learning is given.
出处
《计算机科学》
CSCD
北大核心
1994年第5期27-30,共4页
Computer Science
关键词
神经网络
容限空间理论
扩展
Tolerance spaces theory, Artificia neural network, Machine learning, AI.
