摘要
本文将基于分辨矩阵的二元矩阵和基于属性重要度的启发式属性约简算法结合起来,提出了一种新颖的针对不完备信息系统的属性约简算法。该算法用条件属性和决策属性之间的依赖度来度量属性重要度,进行启发式约简。该算法将属性约简问题转化为寻找能够覆盖决策属性的二元矩阵的二元矩阵集合问题。通过实例检验,该算法是有效的。
A novel attribute reduction algorithm in incomplete information （IIS） system is proposed, combining with binary matrix and attribute-oriented heuristic reduction. The significance of the conditional attributes is defined to measure the ability about determining the dependency relation between the conditional attributes and the decision attributes. The issue of finding an attribute reduction is converted to the issue of searching a set of binary matrices, in which the union of the binary matrices can cover the binary matrix of the decision attribute. Through the illustrative example of car relation, this algorithm is proved to be effective based on the tolerance relation.
出处
《数字技术与应用》
2010年第3期122-124,共3页
Digital Technology & Application
