【摘要】 针对密度峰值聚类算法由于时空复杂度高而不能对大数据集进行有效聚类的问题,提出一种基于网格的密度峰值聚类算法.首先通过自适应多分辨率的网格划分的思想把数据划分到多个网格空间中,然后在每个网格空间中进行密度峰值聚类,利用网格边界合并网格空间中的聚类结果,从而得到原始数据集的聚类结果.本算法集成了网格聚类和密度峰值聚类算法的优点,网格的方法可以减少算法空间复杂度和算法计算量,进而降低了密度峰值聚类算法的时空复杂度.仿真实验结果表明,本算法能够有效处理数据聚类问题,并提高了传统算法的效率.更多还原

【Abstract】 Based on the problem that density peak clustering algorithm cannot cluster large data sets due to the high complexity of time and space effectively,this paper proposed a new density peak clustering algorithm using grid to solve this problem.Firstly,according to the idea of adaptive multi-resolution for grid division to divide data into the grid spaces,performed density peak clustering in each grid space,then through boundary grid combined the clustering results in grid space,to get clustering results of the original data sets.This algorithm integrated the advantages of grid clustering and density peak clustering algorithm,the method of grid division to reduce the time and space complexity,and lead to the same to reduce the time and space complexity of density peak clustering algorithm.The results of simulation showed that,this algorithm can deal with the problem of data clustering effectively,and improve the efficiency of the traditional algorithm.更多还原