不完备决策系统下的多特定类广义决策约简  

The multi-class-specific generalized decision preservation reduction in incomplete decision systems

作　　者：唐玉凯 张楠[1,2] 童向荣 张小峰[3] TANG Yukai;ZHANG Nan;TONG Xiangrong;ZHANG Xiaofeng(Key Lab for Data Science and Intelligence Technology of Shandong Higher Education Institutes,Yantai University,Yantai 264005,China;School of Computer and Control Engineering,Yantai University,Yantai 264005,China;School of Information and Electrical Engineering,Ludong University,Yantai 264025,China)

机构地区：[1]烟台大学数据科学与智能技术山东省高校重点实验室,山东烟台264005 [2]烟台大学计算机与控制工程学院,山东烟台264005 [3]鲁东大学信息与电气工程学院,山东烟台264025

出　　处：《智能系统学报》2019年第6期1199-1208,共10页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金项目(61572418,61572419,61873117,61403329);山东省自然科学基金项目(ZR2018BA004,ZR2016FM42)

摘　　要：属性约简是粗糙集理论研究中最重要的领域之一。经典的不完备决策系统广义决策约简关注决策系统中的所有决策类,而在实际应用中,决策者往往只关注一个或者几个特定决策类。针对以上问题,提出基于多特定类的不完备决策系统广义决策约简理论框架。首先,定义了单特定类的不完备决策系统广义决策约简的相关概念,提出并证明相关定理,构造相应差别矩阵和区分函数。其次,将单特定类的广义决策约简推广到多特定类,提出基于差别矩阵的多特定类的不完备决策系统广义决策约简算法。最后,采用6组UCI数据集进行实验。实验结果表明,相对全部决策类数量,当选定特定类数量较少时,平均约简长度有不同程度的缩短,占用空间有所减小,约简效率有不同程度的提升。Attribute reduction has an important place in rough set theory.The method of classical generalized decision preservation reduction in incomplete decision systems is to find the reducts of all decision classes.In practical applica-tions,however,the decision makers may focus on one or several decision classes.To fill this gap,the theoretical frame-work of multi-class-specific generalized decision preservation reduction in incomplete decision systems is proposed.First,the single-class-specific generalized decision preservation reduction in incomplete decision systems is defined.Re-lated theorems are proposed and proven,and the corresponding discernibility matrix and function are constructed.Then,the single-class-specific generalized decision preservation reduction is extended to the multi-class-specific generalized decision preservation reduction in incomplete decision systems.The algorithm of the multi-class-specific generalized decision preservation reduction based on discernibility matrix in incomplete decision systems(MGDRDM)is proposed.Finally,six datasets from UCI were used for experiments.The experimental results show that when the number of selec-ted specific classes is less than all the decision classes,the average length of reducts will be shortened to varying de-grees,the space used will be reduced,and the time efficiency will be roughly improved.

关 键 词：粗糙集 属性约简 不完备 决策系统 相容关系 多特定类 广义决策约简 差别矩阵 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]