基于容差关系的多粒度粗糙集中近似集动态更新方法

Dynamic updating method of approximations in multigranulation rough sets based on tolerance relation

针对不完备信息系统变化时缺失值获取具体属性值的特性,为解决多粒度粗糙集中更新近似集时间效率低的问题,提出了一种基于容差关系的近似集动态更新算法。首先,讨论了基于容差关系的近似集变化的性质,并根据相关性质得出乐观、悲观多粒度粗糙集的近似集的变化趋势;然后,针对更新容差类效率低的问题,提出了动态更新容差类的定理;最后,在此基础上,设计出基于容差关系的近似集动态更新算法。采用UCI数据库中4个数据集进行仿真实验,当数据集变大时,所提更新算法的计算时间远小于静态更新算法的计算时间,即所提动态更新算法的时间效率高于静态算法,验证了所提算法的正确性和高效性。

Focused on the issue that missing attribute values are obtained when an incomplete information system changes, in order to solve the problem of low time efficiency of updating the approximations in a multigranulation rough sets, a dynamic update algorithm based on tolerance relationship was proposed. Firstly, the properties of the approximations change based on tolerance relationship were discussed, and the change trends of the approximations of optimistic and pessimistic multigranulation rough sets were obtained according to the relevant properties. Then, a theorem of dynamic update tolerance class was proposed for the problem of low efficiency of updating tolerance class. Based on this, a dynamic update algorithm based on tolerance relationship was proposed. The simulation experiments were carried out using four data sets in UCI database. When the data set becomes larger, the calculation time of the proposed update algorithm is much smaller than that of the static update algorithm. The experimental results show that the time efficiency of the proposed dynamic update algorithm is higher than that of the static algorithm, which verifies the correctness and efficiency of the proposed algorithm.