基于压缩决策表的乐观多粒度粗糙集粒度约简算法

粒度约简是多粒度粗糙集研究的一个关键问题。为了从乐观多粒度粗糙集的角度研究粒度约简问题，消除冗余数据，提高粒度约简的效率，提出基于压缩决策表的乐观多粒度粗糙集粒度约简算法。针对乐观多粒度粗糙集模型，引入下近似分布粒度约简的概念；利用线性时间排序算法进行等价类划分，为决策表的压缩和下近似集的计算打下基础；以冗余的决策表为研究对象，以核粒度为初始粒度约简集，以粒度重要性为启发式信息，运用粒度约简算法进行粒度约简，并通过实例分析和实验验证了该算法的有效性。结果表明，算法降低了计算下近似集的时间复杂度，具有较高的粒度约简效率。

Granular space reduction algorithm to optimistic multi-granulation rough sets based on compressed decision table

Granular space reduction is a key issue in the research of multi-granulation rough sets. In order to eliminate redundant data and speed up the granular space reduction from perspective of optimistic multi-granulation rough sets, a granular space reduction algorithm to optimistic multi-granulation rough sets based on compressed decision table is presented. Firstly, a concept of distribution granular space reduction is introduced. Secondly, equivalence classes are divided using linear time sort algorithm for computing compressed decision table and lower approximation. On this basis, granular space reduction is conducted using the proposed granular space reduction algorithm. Finally, the example and experiments are given to demonstrate the validity of the algorithm. The results show that the proposed algorithm can greatly reduce the time complexity for computing lower approximation and has higher granular space reduction efficiency.