中文说明：应用背景 对于K-medoids的原点敏感性问题以及庞大资源群使其功能下降等明显的缺点，学者们研发了一个依靠原点微量调整和增量原点延迟集选的优化 K-medoids 聚类算法。新的算法就是基于微量调整的形式来改善原始点，用原点集合渐渐的展开计算方法以便有效的减少原点替换的运行时间。结论如我们的设想一样，与以前的K-medoids算法比较，我们研究的聚类算法可以显著的提升算法的效率，可以大幅度的减少运行时间。 关键技术 K-medoids属于最简洁的无时空的分析算法,克服了大多数困扰人们的信息处理难题。经过特定的信息库(设有k集群)固定的预先证明，对给定的数据集来划分过程要遵循简洁而且方便的办法。重点的思维是对每个集群定义k质心。由于不一样的位置会得出不一样的结果，这些质心应该放置在一个巧妙的位置上。因此比较好的选择是把它们尽量相互拉开。接下来要做的就是把全部的点归类到一个定义好了的集群中去,并将它与邻近的原点结合在一起。当所有的点都被分配联系完成后,第一步就完成或者提前完成了。
English Description：
Application backgroundFor the origin sensitivity of k-medoids and huge resources that enable the function decline obvious disadvantages, scholars R & D the a rely on origin of micro adjustment and incremental origin delay set selected optimization & nbsp; K-medoids clustering algorithm. The new algorithm is based on the form of trace adjustments to improve the original point, and the origin set gradually developed method of calculation in order to effectively reduce the origin of the replacement of the running time. Conclusions compared with the previous K-medoids algorithm, the clustering algorithm can significantly improve the efficiency of the algorithm, and can greatly reduce the running time.. Key TechnologyK-medoids belongs to the most concise and no time and space analysis algorithm, and overcomes most of the problems that people have to deal with.. After a certain information base (k cluster) is fixed, it is proved that the