一种具有对称性的变精度粗集模型

A Limited and Variable Precision Rough Set Model with Symmetry

为了能够从不完备决策表(IDT)中进行知识发现和数据挖掘,提出一种新的具有对称性的双重可变精度限制容差关系粗集模型(VPLTRST).在该模型中,设定一对可调的参数使其作用于由IDT衍生出的限制容差关系,从而形成上下近似集.文中还提出新的在该模型下的知识依赖以及依赖度的定义,并以此作为依据进行知识约简.结合实例,清晰详实地展示了双重精度下如何获得所有知识约简并最终获得决策规则的全过程,具有很好的效果.

To obtain knowledge and data from incomplete decision table（IDT）, the paper presents a new doubly variable precision limited tolerance rough set theory model （VPLTRST）. The model impacts a pair of tunable parameters on the tolerance relation derived from an IDT between objects by attribute values and a threshold on inclusion relation between subsets to form lower and upper approximations in the model for processing the table. The paper also introduces new definitions of knowledge dependency and dependency degree as basic concepts to implements reducts under this new model. Through examples, under doubly variable precision, it syllabifys and detailedly shows the process how to get all reducts and acquire the decision rules. The model is rational and efficient.