【摘要】 为解决模糊粗糙集属性约简中模糊相似关系运算的不合理性以及多类型数据的不适用性，提出一种基于混合核模糊熵的多类型数据属性约简算法。提出基于核函数的多类型属性模糊关系计算方法，称之为混合核函数度量，并构造出相应的模糊信息粒化模型；利用混合核函数度量进一步提出多类型数据的模糊互补信息熵模型和相关性质；利用模糊互补条件熵和模糊互补互信息熵，提出多类型数据信息系统的不确定性度量和属性约简。实验结果验证了所提出不确定性度量和属性约简方法在多类型数据上的有效性。更多还原

【Abstract】 In fuzzy rough set attribute reduction, a multi-type data attribute reduction algorithm based on mixed kernel fuzzy entropy was proposed to address the irrationality of fuzzy similarity operation and the inapplicability of multi-type data. A kernel based method for calculating multi-type attribute fuzzy relation was proposed, which was called mixed kernel function measurement, and a corresponding fuzzy information granulation model was constructed. The fuzzy complementary information entropy model and related properties of multi-type data were further proposed by using the mixed kernel function measurement. The fuzzy complementary condition entropy and the fuzzy complementary mutual information entropy were used to propose uncertainty measurement and attribute reduction of multi-type data information system. Experimental results validate the effectiveness of the proposed uncertainty measurement and attribute reduction method on multi-type data.更多还原