论文部分内容阅读
将模糊粗糙集作为经典粗糙集的一种模糊推广,把模糊集合相似度引入模糊粗糙集模型中,提出一种基于变相似度的模糊粗糙集模型.通过定义模糊相似矩阵和不一致程度矩阵,给出知识约简的相关概念,即属性约简、核以及属性约简的算法,并证明模糊最小约简和核之间的关系.通过实例证实,对相似精度进行调节,在获得多个层次上的属性归约集的同时保证分类准确性,增加对信息的灵活掌握.最后对本文算法和紧计算域的模糊粗糙集算法进行对比测试,结果验证采用本文算法的约简结果具有较高的分类准确率,为解决粗糙集中连续属性的属性约简问题提供一个有效的方法.
The fuzzy rough set is regarded as a kind of fuzzy generalization of classical rough set, the similarity of fuzzy sets is introduced into the fuzzy rough set model, and a fuzzy rough set model based on variable similarity is proposed. By defining fuzzy similar matrix and inconsistent degree matrix, The related concepts of knowledge reduction, namely, attribute reduction, kernel and attribute reduction algorithms, and prove the relationship between fuzzy minimum reduction and kernel.Through the example validation, the adjustment of similar precision, in obtaining multiple levels The attribute set of attribution is guaranteed, and the accuracy of classification is ensured and the information is flexibly grasped.Finally, the algorithm of this paper is compared with the fuzzy rough set algorithm in the tight computational domain, and the result verifies that the reduction result using this algorithm has a higher classification The accuracy rate provides an effective method to solve the attribute reduction problem of continuous attributes in rough set.