概率复合粗糙集模型的改进及其属性约简

Improved probabilistic composite rough set model and attribute reduction

复合粗糙集及其扩展模型可同时对信息系统中多个二元关系进行有效处理,已在实际中得到成功应用.研究了概率复合粗糙集模型中的复合关系,通过对信息系统中数值类型数据进行处理,并将其转换为区间值类型来简化复合关系,进而提出了一种改进的概率复合粗糙集模型及分布属性约简方法.最后通过实例验证了该方法的合理性.

Attribute reduction is an important task in data mining and the theory of rough set has been applied to attribute reduction successfully.However,in most extended rough sets,there is a supposition that there exists only one data type and one relationship in an information system.So it is difficult to deal with multiple binary relations in an information system.Therefore,some researchers also proposed composite rough set model.The composite rough sets and their extended models can effectively deal with multiple binary relations in information systems and have been successfully applied in practice.In this paper,the composite relation in the probabilistic composite rough set model is studied.By processing the numerical data in the information system and converting it into the intervalvalued data to simplify the composite relation,we propose a novel probabilistic composite rough set model and the corresponding distribution attribute reduction method.Finally,the rationality of the method is verified by an example.