聚类是数据挖掘中用来发现数据分布和隐含模式的一项重要技术。全面总结了数据挖掘中聚类算法的研究现状,分析比较了它们的性能差异和各自存在的优点及问题,并结合多媒体领域的应用需求指出了其今后的发展趋势。
Clustering is an important technique in Data Mining (DM) for the discovery of data distribution and latent data pattern. This paper provides a detailed survey of current clustering algorithms in DM at first, then it makes a comparison among them, illustrates the merits existing in them, and identifies the problems to be solved and the new directions in the future according to the application requirements in multimedia domain.