基于非平衡数据下不完备混合型信息系统的属性约简

完备混合型信息系统下的粗糙集模型是传统粗糙集模型的重要扩展,目前关于非平衡数据属性约简的研究仅限于完备混合型的粗糙集模型。针对这一问题,提出一种基于不完备混合型信息系统的非平衡数据属性约简。本文首先将传统的粗糙集模型进行推广,提出不完备混合型信息系统下的粗糙集模型;然后针对数据的非平衡性,根据上下边界区域和类分布的不均匀性定义了一种新的属性重要度;在基于区别矩阵的基础上设计出一种非平衡数据的属性约简算法。实验分析表明该算法针对不完备非平衡数据的属性约简具有一定的有效性和优越性。

Attribute reduction of incomplete hybrid information system based on unbalanced data

The rough set model under complete hybrid information system is an important extension of the traditional rough set model.However,the research on the attribute reduction of unbalanced data is limited to the complete hybrid rough set model.To solve this problem,this paper proposed an unbalanced data attribute reduction based on incomplete hybrid information system.This paper generalized the traditional rough set model,and then proposed the rough set model under incomplete hybrid information system.Then,it defined a new attribute importance according to the heterogeneity of upper and lower boundary regions and class distribution for the unbalanced data.Finally,this paper designed an algorithm for attribute reduction of unba-lanced data based on discernible matrix.The experimental results show that the algorithm is effective and advantageous for attribute reduction of incomplete and unbalanced data.