不完备信息系统下的变精度粗糙集模型及其知识约简算法

1 引言 Z.Pawlak提出的粗糙集理论,提供了一种新的知识获取手段,但是,应用经典粗糙集理论的一个重要前提是它的处理对象必须是完备信息系统。现实中,由于数据采集能力不足等原因又使得大量的信息系统是不完备的,因此需要对经典粗糙集理论进行必要的扩充。目前有两种扩充方法,一是间接处理方法,即先通过一种方式将不完备信息系统转化为完备信息系统,再使用针对完备信息系统的方法对其进行处理;二

Variable Precision Rough Set Model and a Knowledge Reduction Algorithm for Incomplete Information System

Rough Set Theory, which has been found applicable and useful in many fields, is now a very effective method in data mining research. However, when the decision table is an incomplete one, with the original rough set theory proposed by Z. Pawlak, one can't get satisfactory results. In this paper an approach based on limited valued tolerance relation and majority inclusion relation is proposed. And furthermore a new attribute reduction method called extended discernable matrix is given. As this model is somewhat a combination of fuzzy means and majority inclusion relation, it is more effective than the previous models in practice.