针对不一致决策表的规则获取,从属性多粒度角度考虑,按粒度由粗到细将决策表划分成不同的粒度空间,通过定义相容粒关系矩阵和不相容粒关系矩阵,并充分挖掘隐含在矩阵中的启发式信息,实现对不同粒度空间确定性规则和不确定性规则的获取.最后,从实例分析以及UCI测试对算法进行验证,并与现有算法进行实验对比,实验结果与分析表明了所提出算法的可行性和有效性,而且按此方法获取的规则集的泛化能力更强. According to the multi-granularity of attribute, the decision table is divided into different granularity spaces according to the granularity from coarse to fine. By defining the matrix of compatible grain relation and the matrix of incompatible grain, The heuristic information contained in the matrix realizes the obtaining of the rules and uncertainty rules of different granularity space.Finally, the algorithm is verified from the example analysis and the UCI test, and compared with the existing algorithms, the experimental results and The analysis shows the feasibility and validity of the proposed algorithm, and the rule set obtained by this method has more generalization ability.