摘要
分类是数据挖掘中一个重要的研究领域。针对原始决策表中往往存在大量冗余信息,从而影响决策分类综合性能这一问题,提出了一种基于粗糙集和RBF神经网络的分类模型。该模型在保持训练样本分类质量的情况下,运用属性约简方法对决策表进行约简,得到维数较小的训练样本空间。通过这样确定RBF神经网络输入层变量,优化了网络结构。实例结果表明了该方法的有效性和实用性。
Classification is an important research field in data mining.It was found that the precision and speed of decision is unsatisfied due to large-scale redundant information in decision table.To solve the problem,a new classfication model based on rough set and RBF neural network is presented.. In order to set input variables of the RBF neural network,attribute reduction algorithm is used firstly to reduce the dimension of the vector space by still maintaining classification quality of training samples,and opt...
出处
《宿州学院学报》
2008年第4期103-105,共3页
Journal of Suzhou University
基金
安徽师范大学专项基金资助项目(2005Bzx19)
安徽省教育厅自然科学基金项目(2005KJ094)
关键词
粗糙集
属性约简
RBF神经网络
分类模型
Rough sets
Attribute reduction
RBF neural network
Classfication model
