基于粗糙集属性重要性的数据约简

为了提高数据处理的效率,本文以等价关系、划分、上下近似、正域等粗糙集理论基本概念为基础,提出基于属性重要性的属性约简方法。该方法从条件属性集整体出发,分别计算去除每个条件属性后对决策属性重要性的变化,利用属性重要性区分不同属性对于决策属性的依赖程度,优选重要程度较高的属性。该方法以工业现场交流调速系统数据为对象进行有效性验证,结果表明,该方法具有良好的性能,能够大量减少冗余的属性和数据,提高数据处理的效率。

Data reduction based on importance of attribute

In order to improve the efficiency of data processing,this paper figures out a reductive method based on rough set theory related conceptions,such as equivalent relationship,upper/nether approximate,and positive region.The method originates from conditional attributes;calculates the difference of importance of decision attribute after taking out one conditional attribute each time separately.And the dependency is classified based on the importance of attributes,important attributes are selected.The method has been test on alternating current governor system;it can improve the performance of data processing.