# cluster by fast search and find of density peaks

This paper proposed a new cluster idea. The idea is that the cluster center is characterrized by a higher density than their neighbors and by a relatively large distance from points with highter density(1.一个类中的聚类中心的点的密度较高，2.不同聚类中心的距离较大).

Based on this assumption, for each data point , we compute two quantities: Its local density  and its distane  from point with higher density. Both quantities are based on this distance  .The local density  is defined as

where  if  <0 and  otherwise, and  is a cutshort distance.(影响变量有 ).

is measured by computing the minimum distance between the point and any other points with a higher density,(在密度比它大的数据点钟寻找距离最小的点).

http://www.sciencemag.org/content/344/6191/1492.full

## Science论文&quot;Clustering by fast search and find of density peaks&quot;学习笔记

"Clustering by fast search and find of density peaks"是今年6月份在<Science>期刊上发表的的一篇论文,论文中提出了一种非常巧妙的聚类算法.经过几天的努力,终于用python实现了文中的算法,下面与大家分享一下自己对算法的理解及实现过程中遇到的问题和解决办法. 首先,该算法是基于这样的假设:类簇中心被具有较低局部密度的邻居点包围,且与具有更高密度的任何点有相对较大的距离.对于每一个数据点,要计算两个量:点的局部密度和

## Clustering by fast search and find of density peaks

"Clustering by fast search and find of density peaks"是20114年6月份在<Science>期刊上发表的的一篇论文,论文中提出了一种非常巧妙的聚类算法. 首先,该算法是基于这样的假设: (1)聚类中心密度要高 (2)高密度中心点之间的距离应该相对远一些. 异常点都会被排除,同时也和形状无关. 首先,这种方法不像原先的Kmeans那样随机初始种子点然后迭代,它是根据样本的密度峰值来确定聚类中心的,当然聚类中心确定之后,后面