摘要
基于粗糙集理论,对基于属性重要度的启发式属性约简算法进行了改进,并将改进后算法运用于纳税人属性约简的实际工作中.该算法解决了原有基于属性重要度的启发式属性约简算法结果中存在冗余属性问题,实现了属性选择较小化,并保持原有数据分类能力不发生大的变化.通过属性约简实验结果和实际工作情况对比,证明该算法具有很好的性能.
In reference to rough set theories, this paper improves the heuristic attribute reduction method based on importance degree ,and applies to the practice of the real tax of taxpyer＇s attribute reduction. Through this method, the rebandant attribute of heuristic attribute reduction method based on importance degree, is solved. This method cam minimize the attribute selection, and keep the ability to classify tax data without great changes. Compared with the tested result of attribute reduction and the real work, it proves that this method is effective.
出处
《微电子学与计算机》
CSCD
北大核心
2008年第8期212-215,共4页
Microelectronics & Computer
关键词
粗糙集
启发式
税源分析
属性选择
属性约简
rough sets
heuristic
tax source analysis
attribute selection
attribute reduction
