摘要
在数据挖掘研究过程中,对连续型属性一般要进行离散化。特别是在模糊数据挖掘中,还要对离散化的区间进行模糊处理。文中依托云模式,并结合粗糙集理论提出一种新的连续型属性离散化算法。
In the process of data mining, it is necessary to discretize the continuous-valued attributes. In the process of fuzzy data mining, it is also necessary to make the intervals of discretization fuzzy. This paper gives a new discretization algorithm of continuous-valued attributes based on cloud models together with the theory of rough sets. After a set of cuts is decided based on rough set theory, a membership cloud is constructed on each cut by maximum-likelihood method, and the membership clouds are mergered according to subsethood measure at last. Borders of the concepts produced by this method are fuzzy, and the inexactness of the original information system can be kept .
出处
《计算机应用》
CSCD
北大核心
2004年第2期135-137,共3页
journal of Computer Applications
关键词
模糊数据挖掘
离散化
隶属云
包含度
fuzzy data mining
discretization
membership clouds
subsethood measure