一种否定高信度规则的属性查找算法

粗糙集理论中可辨识矩阵是在整个论域U上构造的。该文扩展了可辨识矩阵的应用，提出一种否定高信度决策规则的属性查找算法，求出条件属性和决策属性的不可分辨二元关系在等价类上进行运算。在论域U的子集上构造可辨识矩阵，根据分辨函数求解问题。将算法应用于医学数据。实验结果表明，该算法局部采用可辨识矩阵可以有效地减少存储空间，提高查找效率。

Algorithm to Find Attributes Negated High Reliability Decision Rules

In rough set theory, the discernibility matrix is used on the whole data set. In this paper, the application of extended discernibility matrix is presented. The discernibility matrix can be used to not only the whole data set but also the partial data. An algorithm is introduced, which is used to find the attributes that negate the high reliability decision rules on some condition attributes. It selects anyone among all the condition attributes and decision attributes to calculate its indiscernibility binary relation. The algorithm is based on equivalence classes and uses a part of data of the whole data set to construct the discernibility matrix. The question can be solved by the discernibility function. An experiment with medical data is presented. Experimental results show that the algorithm is more efficient and has less space complexity and time complexity.