基于K-均值聚类的混合聚类算法

Hybrid Clustering Algorithm Based on the K-Means Clustering

K-均值聚类算法是聚类算法中比较典型的算法之一,在其各类改进算法中都受到了离群点、初质心、类个数等因素的干扰。本文利用相似密度提出一种新的聚类初始质心选取和离群点判别方法,对K-均值聚类算法进行了改进。通过实验证明改进算法提高了聚类的有效性和稳定性。

K-means is one of typical clustering algorithm,but it is easily affected by the outliers,the initial cluster center of mass,and the given cluster number and so on. In this paper,the method of similar density is used to determine the initial cluster center of mass,then to remove outliers through the outlier determine rules. Furthermore,we make an improvement on K-means clustering algorithm and improve effectiveness and stability of the clustering result by the experiment.