广义模糊粗糙集在模糊信息约简中的应用

针对数据集为模糊值时冗余信息难于消除的问题，提出了基于模糊相似关系的广义模糊粗糙集与QuickReduct算法相结合的方法。利用广义模糊粗糙集数据相似程度对属性值为实数值的数据集合进行约简，不需要预先对原始数据集合进行离散化，约简结果能更完整地反映原信息系统的分类能力。同时算法中利用了启发式信息，使模糊依赖性增加较快的属性作为最小约简。计算实例验证了该方法的有效性。

Generalized fuzzy rough set approach to fuzzy information reduct

The main obstacle facing current dataset of fuzzy value is that of redundancy-removing.The paper proposes the generalized fuzzy rough set based on similarity relation which unifies with"QuickReduct" algorithm.The generalized fuzzy rough set carries on the reduction of attributes by adopting the data set of similar degree to the real value.The quantization of the original dataset is unnecessary beforehand and the ability of classify can not destroyed by reducing attributes.Simultaneously,the heuristic information is adopted in the algorithm and a minimum reduct can be found according to the attribute of fuzzy dependence increasing quickly.It is available by computing and analyzing with fuzzy information.