基于共现概率的三支聚类模型

Three-way clustering model based on co-occurrence probability

三支聚类对不确定对象引入了边界域,可以有效解决传统二支聚类方法中由于信息不完整而导致划分不准确的问题。如何获得三支聚类的核心域和边界域是目前研究三支聚类的重点之一。该文将共现概率与三支聚类相结合,提出了基于共现概率的三支聚类模型。首先,基于朴素贝叶斯确定两样本的共现概率;其次,给出了基于共现概率的相似关系及其粗糙集的下、上近似,获得三支聚类的核心域和边界域;最后,在UCI数据集上的实验结果显示,该方法提高了聚类精度,验证了其可行性。

Three-way clustering introduces the boundary region to uncertain objects,which can effectively solve the problem of inaccurate division caused by incomplete information in traditional two-way clustering.How to obtain the core region and boundary region of three-way clustering is one of the key points of three-way clustering.Combining co-occurrence probability with three-way clustering,a three-way clustering model is proposed based on the co-occurrence probability.Firstly,the co-occurrence probability of two samples is determined based on naive Bayes.The similarity relation,and the lower and upper approximation of rough set based on the co-occurrence probability are then given.The corresponding core and boundary regions of the three-way clustering can be obtained.Finally,the experimental results on UCI datasets show that the proposed method can improve the accuracy of clustering,and the feasibility is also verified.