摘要
从灰关联分析中最核心的灰关联度构造及相应的挖掘算法出发,结合正态分布的普适性,提出了一种体现数据分布特点的正态灰数,并给出了相应的灰度及灰关联度计算方法。在此基础上,构建了一种多粒度无监督的快速灰聚类方法,无需先验知识即可完成自动聚类。通过实验验证了本文方法的有效性,为大数据下灰关联分析的进一步发展提供了新思路。
The classic grey theory does not adequately take into account the distribution of data set, and lacks effective methods to analyze and mine large sample in multi-granularity. Considering the universality of normal distribution, a normality grey number is proposed. Moreover, the corresponding definition and calculation method of the incidence degree between the normality grey numbers are constructed. On this basis, the grey incidence analysis method in multi-granularity is put forward to realize the automatic clustering in the specified granularity without any experience knowledge. Experiments fully demonstrate that the proposed method is effective in knowledge acquisition for large data.
出处
《吉林大学学报（工学版）》
EI
CAS
CSCD
北大核心
2015年第1期283-290,共8页
Journal of Jilin University:Engineering and Technology Edition
基金
国家自然科学基金项目(61309014)
重庆市基础与前沿研究计划项目(cstc2013jcyj A40009
cstc2013jcyj A40063)
重庆市教委科学技术研究项目(KJ1400412)
关键词
计算机应用
灰理论
正态灰数
灰关联度
灰关联分析
computer application
grey theory
normal grey number
degree of grey incidence
grey incidence analysis
作者简介
代劲（1978-），男，副教授，博士．研究方向：智能信息处理，灰理论．E—mail：daijin@cqupt．edu．cn
