优势粗集中的属性约简和对象排序  

Attribute reduction and ordering objects dominance-based rough set

作　　者：张卉[1] 李续武[1] 翟兴隆[1] 

机构地区：[1]空军工程大学防空反导学院,陕西西安710038

出　　处：《计算机工程与设计》2013年第9期3241-3244,共4页Computer Engineering and Design

基　　金：陕西省自然科学基金项目(2006F18)

摘　　要：为进一步解决优势粗糙集中存在的属性约简和对象排序的问题,通过分析已有优势粗糙集模型的共性,提出了基于由粗糙边界引起的不确定概率的属性约简算法和约简对象集后的基于区分矩阵的属性约简算法,并对两种算法的时间复杂度做出了分析,同时通过实例验证了两种算法的实用性和高效性,尤其是第二种算法在处理数据量大信息系统时体现出更大的优势。对于对象排序,提出一种基于优势差的二元对象排序方法,为对象排序提供了一种新的思路。To better solve the two major problems exist in dominance-based rough set approach： attribute reduction and ordering objects, two ways of attribute reduction： attribute reduction based on uncertain probability caused by rough sidewall and attrib- ute reduction on discernibility after reduction on object set by analyzing the universality of dominance-based rough set models which exist and analyzed these two＇s time complexity. At the same time these two arithmetics＇ practicability and high efficiency and the second＇s advantages are verified when large amounts of data are processed by experiment. For ordering objects, a dualis- tic method to ordering objects is proposed, and a new way of ordering objects for the future is provided.

关 键 词：单值完备信息系统 不确定概率 对象集约简 二元对象排序 优势差 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]