论文部分内容阅读
利用粗糙集理论,从矩阵分析的角度来挖掘决策表蕴含的信息,引入粗糙集信息等价关系的同构映射——等价矩阵,等价矩阵可看作是等价关系在信息表内的知识表达。给出了等价矩阵的求取算法以及等价矩阵意义下的属性重要度和核的概念。设计了基于等价矩阵的决策信息表的最小属性约简算法。从等价矩阵本身相关操作运算来挖掘客观知识之间的关联模式,提出了基于信息等价矩阵的关联规则提取的算法。实例证明提出的算法有效,为进一步研究决策信息系统的规则提取和决策算法提供了可行的计算方法。
Using the rough set theory, the information contained in the decision table is extracted from the perspective of matrix analysis. The isomorphic mapping equivalence matrix of the equivalence relation of rough set information is introduced. The equivalence matrix can be regarded as the equivalence relation in the information table Knowledge expression. The algorithm of obtaining equivalent matrix and the importance of attribute and the concept of kernel under the equivalent matrix are given. The minimum attribute reduction algorithm of decision information table based on equivalent matrix is designed. Based on the correlation matrices of equivalent matrices, the association patterns of objective knowledge are tapped and the algorithm of association rules extraction based on information equivalent matrix is proposed. The example proves that the proposed algorithm is effective and provides a feasible method for further studying the rules extraction and decision algorithm of decision information system.