基于网格和密度的聚类算法的分析与研究

Research and Analysis of Clustering Algorithm Based on Grid and Density

针对CLIQUE算法的特点以及所存在的问题进行深入的研究。为了进一步提高其处理高维海量数据的能力,在原算法的基础上提出一种基于密度样本分析和基于最优区间分割进行改进的聚类算法,并通过使用仿真数据加以验证是可行的,理论分析与实验结果表明,与原算法相比,改进算法不仅保留原算法的优点,且对大规模数据集有着很好的聚类效果。

The characters and existing problems of CLIQUE clustering algorithm are intensive researched. In order to improve the ability of solving the high dimention and mass data, based on the old algorithm,a modified one with the methods of density and the best space division is presented. Proving it with simulation data and it is feasible. Theory analysis and experimental results demonstrate the improved algorithm not only can keep its old advantages but also can get better clustering resuits.