具有连续属性的不完备信息系统Rough集扩展

The Extension of Rough Sets in Incomplete Information Systems Containing Continuous Attributes

在模糊相似关系的基础上,针对具有连续属性的不完备信息系统提出了Rough集的扩展模型;利用模糊区间数来表示不完备信息表中缺失的连续属性值,并根据属性值的概率分布情况计算模糊区间数的可能度,在模糊关系的基础上给出了不完备信息表的上近似和下近似的定义;利用基于欧氏距离的贴近度法计算模糊相似度,构造相似矩阵,实现了对论域的划分;给出决策规则的约简和表示方法.应用实例说明了所提出的Rough集模型及规则的实用性.

A rough set model based on fuzzy similarity relation for incomplete information systems containing continuous attributes was extended. Values of missing continuous attributes in incomplete informational table were represented by interval valued fuzzy numbers, and the probability was calculated based on the probability distribution. The definition of upper and lower approximation for the incomplete information system was given based on the fuzzy similarity relation. Fuzzy similarity ratio based on Euclidean distance was calculated to construct the fuzzy similarity matrix and the universe classification was realized. The method of decision rules reduction and representation was also presented. At last an example was given for further illustration.