摘要
针对经典粗糙集模型在处理不完备、动态数据方面的不足,通过分析容差关系模型,引入先验概率在知识估计中的方法,给出了一种基于区分矩阵的增量式属性约简算法.以属性重要度为启发信息,对区分矩阵的构造过程进行改进,仅需简单的矩阵运算就可以得到约简结果.最后通过示例分析处理增量式数据的算法复杂度有效,算法正确可行.
Aiming at the disadvantage of classical rough set model in dealing with imcomplete and dynamic data,a prior probability method in estimating knowledge is introduced through analyzing tolerance relation,and a incremental attribute reduction algorithm based on discernibility matrix is proposed.Using attribute significance as heuristic message,the process of constructing discernibility matrix is improved,reduction result can be got only by simple matrix computing.Finally,algorithm's complexity in dealing with incremental data is effective through example analysis,and the algorithm is valid and feasible.
出处
《辽宁工程技术大学学报（自然科学版）》
CAS
北大核心
2012年第2期284-288,共5页
Journal of Liaoning Technical University (Natural Science)
基金
海南省自然科学基金资助项目(610221)
海南师范大学青年教师科研启动基金资助项目(QN0918)
关键词
粗糙集
不完备系统
增量式约简
区分矩阵
属性重要度
先验概率
容差关系
算法复杂度
rough set
imcomplete system
incremental reduction
discernibility matrix
attribute significance
prior probability method
tolerance relation
algorithm's complexity
