This paper has proposed a classification rule base mining algorithm combining the genetic algorithm and intuitionistic fuzzy-rough set for large-scale intuitionistic fuzzy information system. The algorithm has proposed innovatively the definitions and measurement metrics of completeness, interaction and compatibility describing the whole rule base, and constructed a multi-objective optimization model to optimize the population size of data sample, and used the intuitionistic fuzzy-rough set to reduce the attribute set of fuzzy information system, and used intuitionistic fuzzy similar class to extract large-scale intuitionistic fuzzy information system rules, and obtained an optimal rule base with the minimal size, configuration, generation time and storage space. A threshold control mechanism is used to evaluate the completeness, interaction and correlation of rule population and improves the robustness and flexibility of sample population optimization and rule base generation. The algorithm is verified by the real aircraft health data sets. The algorithm is validated by similar mature and effective algorithms in accuracy, time complexity using real data and has good robustness and adaptability to different size large-scale fuzzy information system.

中文翻译：

---

直觉模糊粗糙集与遗传算法相结合的分类规则挖掘算法

针对大规模直觉模糊信息系统，提出了一种结合遗传算法和直觉模糊粗糙集的分类规则库挖掘算法。该算法创新性地提出了描述整个规则库的完整性，交互性和兼容性的定义和度量标准，并构建了一个多目标优化模型来优化数据样本的总体大小，并使用直觉模糊粗糙集来减少模糊信息系统的属性集，并使用直觉模糊相似类来提取大规模直觉模糊信息系统规则，从而获得具有最小尺寸，配置，生成时间和存储空间的最优规则库。阈值控制机制用于评估完整性，规则总体的交互作用和相关性，并提高了样本总体优化和规则库生成的鲁棒性和灵活性。该算法已通过实际飞机健康数据集验证。该算法已通过类似成熟有效的算法在真实性，准确性，时间复杂度等方面进行了验证，对不同规模的大型模糊信息系统具有良好的鲁棒性和适应性。