Abstract: In real-life applications, an information system may vary with time. This paper discusses the approaches for dynamically acquiring approximations in variable precision rough set model while adding or deleting an attribute respectively in information system. By dividing original equivalent classes in information system, this paper proposes an approach which avoids the re-division of the universe and improves the efficiency of dynamically acquiring approximations. By discussing the relationship between equivalent classes and original approximations, this paper gives the corresponding theorems between updated approximations and original approximations, and proposes the approaches for dynamically acquiring approximations while adding or deleting an attribute in variable precision rough set model respectively. The experimental results verify the validity of the proposed approaches, and the efficiency of the proposed approaches is better than that of the original approaches.

Key words: knowledge discovery, approximation, dynamic acquisition

摘要： 在实际应用中，信息系统随着时间在不断发生变化。分别讨论了信息系统中属性增加和减少时，变精度粗糙集模型中近似集的动态获取方法。通过对信息系统中原有的等价类进行划分，避免了对论域的重新划分，提高了动态获取近似集的效率；通过讨论等价类与原有近似集之间的关系，给出了信息系统动态更新之后的近似集与原来近似集之间的相关定理，提出了在变精度粗糙集模型中属性增减时近似集动态获取方法。实验结果验证了该方法的有效性，而且效率优于原始方法。

关键词: 知识发现, 近似集, 动态获取