由于粗糙集只能对离散属性进行处理,因而连续属性的离散化也就成了粗糙集的主要问题之一。提出了一种从模糊聚类出发的离散化方法,并给出了一个判别函数,由该函数从聚类结果中选择最优的一个解,因而是一种自寻优的求解过程,避免了人为划分类数的主观影响。最后进行了实验比较,证实了该方法的有效性和合理性。
Because traditional rough set theory can only deal with the discrete attributes in database. So, the discretization of continuous attributes is one of the main problems in rough sets. In this paper, a method of dlseretization by fuzzy cluster is offered and a criterion function is given, it can select the best solution from the cluster, so it is a superior solution process and subjective influence is avoided. Finally, some experiments are applied to compare with others. The result proved this method is effective.