基于密度偏差抽样的聚类算法研究

随着网络的普及和信息量的急剧增加，从海量数据中提取有用的数据信息已迫在眉睫。在对已有的基于密度偏差抽样算法改进的基础上，提出了一种基于密度偏差抽样的聚类算法。实验表明，随着信息量.数据维数的增加，该算法聚类的正确率以及对数据的处理速度都要较传统的聚类算法有所提高。

Research on clustering algorithm based on density biased sampling

As the popularization of the internet and the quick increase of information, it is more difficult to find the useful knowledge from huge data sets. On the foundation of the existing algorithm based on density biased sampiing, a new clustering algorithm by density biased sampling is proposed. The experimental results show that the proposed algorithm is indeed better than the traditional algorithm as the increase of information and data dimensions, such as the clustering correctness and the sDeed for dealing with data.