一种基于密度和网格的簇心可确定聚类算法

以网格化数据集来减少聚类过程中的计算复杂度,提出一种基于密度和网格的簇心可确定聚类算法.首先网格化数据集空间,以落在单位网格对象里的数据点数表示该网格对象的密度值,以该网格到更高密度网格对象的最近距离作为该网格的距离值;然后根据簇心网格对象同时拥有较高的密度和较大的距离值的特征,确定簇心网格对象,再通过一种基于密度的划分方式完成聚类;最后,在多个数据集上对所提出算法与一些现有聚类算法进行聚类准确性与执行时间的对比实验,验证了所提出算法具有较高的聚类准确性和较快的执行速度.

A density-based and grid-based cluster centers determination clustering algorithm

A density and grid based cluster centers determination clustering algorithm is proposed. The computational complexity of the clustering process is reduced by using the gridding dataset. Firstly, the dataset space is divided into grids with the same size, and the number of data objects that are contained in grid is defined as the value of the grid density. The nearest distance from one grid to another with higher density is defined as the value of grid distance. The cluster center grids can be found since these grids always have high density value and large distance value. Then, a density-based division approach is used to accomplish the task of clustering. Finally, a comprehensive comparison is carried out to examine the clustering accuracy and execution time between the proposed clustering algorithm and some classical algorithms. Experiment results show that the proposed algorithm can lead to a higher accuracy with less execution time.