一种基于改进的ROUSTIDA算法的数据补齐方法  

Improved method for data reinforcement based on ROUSTIDA

作　　者：田树新[1,2] 吴晓平[1] 王红霞[1] 

机构地区：[1]海军工程大学电子工程学院 [2]91829部队

出　　处：《海军工程大学学报》2011年第5期11-15,共5页Journal of Naval University of Engineering

基　　金：国家自然科学基金资助项目(60774029)

摘　　要：针对ROUSTIDA数据补齐方法存在的问题,提出了一种改进的ROUSTIDA数据补齐方法。该方法首先将条件属性与决策属性区分对待,优先填补决策属性值,避免了不一致决策表的产生;然后,在决策属性相同的前提下填补条件属性缺失值,扩展了原算法的使用范围;最后,通过一个实例说明改进算法补齐后的决策表的确能够获得更集中的决策规则。For the deficiency in ROUSTIDA, an improved method was proposed. Firstly, in dealing with the condition attributes and decision attributes in different ways, priority was given to the decision attributes to avoid generating the inconsistent decision system. Then the missing condition attribute values were filled under the prerequisite of the same decision attribute value to expand the application of the original algorithm. Finally, an example was used to show that the improved method can get more centralized decision rules.

关 键 词：粗糙集 数据补齐 ROUSTIDA 决策属性 条件属性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]