Abstract: Traditional decision-theoretic rough set lacks the ability to deal with various types of complexity data. To tackle this issue, by considering fusion technology, this paper firstly proposes the quantitative composite relation, and further verifies that it is the popularization form of intersection-composite relation and union-composite relation. Subsequently, the corresponding decision-theoretic rough set model is presented. Finally, the fusion process with composite relation is intuitively described by matrix operations, and a novel method is introduced to compute the decision-theoretic rough approximations. The experimental results on UCI datasets show that the model and approach are efficient and stabilized.

Key words: complexity data, decision-theoretic rough set, quantitative composite relation, matrix operations

摘要： 经典决策粗糙集缺乏处理多种类型共存的复杂数据的能力。为解决该问题，首先通过融合的技术手段提出量化复合关系，并验证了量化复合关系是交-复合关系和并-复合关系的推广形式，然后在此基础上构建了相应的决策粗糙集模型。最后借助矩阵运算直观地描述了复合关系的融合过程，并给出了求解决策粗糙近似集的新方法。在UCI数据集上的实验结果表明该模型和方法有效且稳定。

关键词: 复杂数据, 决策粗糙集, 量化复合关系, 矩阵运算