基于密度峰值优化的谱聚类算法

针对经典谱聚类算法无法自适应确定聚类数目、以及在处理大数据量的聚类问题时效率不高的问题,本文提出了一种基于密度峰值优化的谱聚类算法。该方法首先计算数据对象的局部密度,以及每个数据对象与较其他数据对象的最小距离,并依据一定的规则自适应产生初始聚类中心,确定聚类数目。其次,使用Nystr?m抽样来降低特征分解的计算复杂度以达到提高谱聚类算法的效率。实验结果表明,本文方法能够准确地得到聚类数目,并且有效提高聚类的准确率和效率。

Spectral clustering based on density peak value optimization

To deal with the problem that classical spectral clustering algorithm can not automatically determine the number of clusters, and settle large amount of data with low efficiency. This paper proposes a spectral clustering algorithm based on density peak value optimization. The method firstly calculates the local density of data object and minimum distance between each data object and other data objects. Adaptive clustering algorithm is generated to determine the number of clusters and to optimize the number of clusters according to certain rules. Secondly, appending Nystrom sampling can reduce the characteristic decomposition time complexity and improve the efficiency of the algorithm. The experimental results show that this method can accurately obtain the number of clusters and effectively improve the accuracy and efficiency of clustering.