摘要
运用粗集理论的不可分辨关系和不可分辨类的概念和归约计算方法,对原始数据进行精简、概略,然后建立神经网络模型和确定各隐层节点之间的连接关系,使得网络从一开始就具有良好的拓扑结构。这种基于粗集理论的神经网络模型和学习算法具有学习速度快、容错能力强等特点。
This paper uses the concept indistinguishing relation and indistinguishing class, reduction calculate method to reduce the stack data. Then creates the neural network and gets the connection of hide node. The method makes the good structure at the beginning. This algorithm based on rough sets and neural network has character of high speed studying and power tolerance.
出处
《安徽建筑工业学院学报（自然科学版）》
2005年第2期81-84,共4页
Journal of Anhui Institute of Architecture(Natural Science)
