摘要
文章对多属性且属性值为连续的决策系统进行预测,提出了灰粗糙支持向量机预测方法。首先采用灰色关联分析计算出条件属性相对于决策属性的重要度;并对连续属性进行离散化,结合Pawlak属性重要度与灰关联度进行约简;将约简后的条件属性作为影响因子,基于支持向量机对决策属性进行预测。实验结果表明,该方法是有效可行的。
Aimed at the prediction of decision system with multi-attribute and continuous valued attrib- utes, a prediction method of grey rough support vector machine is proposed. Firstly, the significance of condition attribute relative to decision attribute is computed by using grey correlation analysis. Sec- ondly, attributes are reduced by using Pawlak＇ s attribute significance and grey correlation degree after continuous attributes are discretized. Finally, the reduced condition attributes are used as impact fac- tors to predict decision attribute based on support vector machine. The experimental results show that the method is valid and feasible.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
北大核心
2012年第12期1652-1654,1724,共4页
Journal of Hefei University of Technology：Natural Science
基金
福建省教育厅A类科技资助项目(JA12220)
关键词
灰关联分析
粗糙集
支持向量机
多属性
grey correlation analysis
rough set
support vector machine
multi-attribute
作者简介
王晨曦（1981-），女，黑龙江海林人，漳州职业技术学院讲师
