基于分层聚类的k-means算法

为了更好地实现聚类,在分析分层聚类和k-means算法优缺点的基础上提出了一种改进的聚类算法.改进算法将分层聚类和k-means聚类算法的优点相结合,首先采用分层聚类,得到一个初始的聚类结果,然后应用k-means聚类算法继续聚类.实验结果表明,改进算法较原先传统的聚类算法,不但算法执行速度快、效率高,而且聚类效果也比较好。

A Hybrid Hierarchical k-means Clustering Algorithm

In order to obtain better clustering results,after analyzing the advantages and disadvantages of hierarchical and k-means clustering algorithms,a new algorithm which combines the advantages of hierarchical and k-means clustering algorithms is proposed.In the algorithm,hierarchical clustering is carried out at first to get an initial clustering in the first round and then the k-means clustering is carried out in another round.The results of experiment suggest that this new method has faster speed,higher efficiency and better clustering results than previous traditional clustering algorithms.