¹ NanChang Institute of Technology, Nanchang, Jiangxi 330099, PR China; Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong 518055, PR China
² Key laboratory of intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100080, PR China
³ Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong 518055, PR China

---

Abstract—Attribute reduction is one of the core research content of Rough sets theory. Many existing algorithms mainly are aimed at the reduction of consistency decision table, and very little work has been done for attribute reduction aimed at inconsistency decision table. In fact, the methods finding Pawlak reduction from consistent decision table are not suitable for inconsistency decision table. In this paper, we introduce the approximation dependency reduction modal and present the Quick Attribution Reduction based on Approximation Dependency Degree (Quick-ARADD), which can retain the original boundary region and the original positive region unchanged, and keep the approximation accuracy unchanged for all decision equivalence classes (the partition of universe on decision attributes) of a decision table. Theoretical analysis and experimental results show that the Quick-ARADD algorithm is effective and feasible.

Index Terms—Rough Set, Attribute Reduction, Approximation Precision, Approximation Dependency, Classification

[PDF]

Cite: Min Li, ShaoBo Deng, Shengzhong Feng, and Jianping Fan, " Quick Attribute Reduction Based on Approximation Dependency Degree," Journal of Computers vol. 8, no. 4, pp. 920-928, 2013.