区间值优势关系系统中属性约简及规则提取方法

Attribute Reduction and Rule Extraction Methods in Interval-valued Dominance Relation Systems

本文以包含偏序关系的区间值决策系统为研究对象,对连续属性值进行模糊化处理,构造一种模糊优势关系粗糙集模型,并讨论了其相关性质。基于新模型提出一种不确定性度量-模糊粗糙熵,并以此为启发信息构造一种启发式约简算法,同时给出了该算法的时间复杂度分析结果。由该算法所得到的决策规则集具备较高的准确度和覆盖度,从而保证了数据预测、分类的准确性和合理性。通过实例分析,证明该算法是区间值优势关系系统中规则获取的有效方法。

In this paper,we fuzzify the continuous attribute values in the Interval-valued decision system with partial-order relations Firstly,a fuzzy dominance relation rough set model is presented and its relevant properties are discussed.Then,an uncertainty measure called fuzzy rough entropy is proposed based on the new model,and this is used as heuristic information to construct a heuristic reduction algorithm.The results of the time complexity analysis of the algorithm are also given.There are high certainty and coverage in the decision rule sets obtained by the algorithm,thus ensuring the accuracy and rationality of data prediction and data classification.Finally,the algorithm is shown to be an effective method for rule acquisition in interval-valued dominance relation systems through an example analysis。