一种基于局部信息的聚类密度度量

为有效处理密度不均匀聚类问题,以数据集蕴涵的局部信息为出发点,提出一种数据点密度度量———松散度,用以揭示数据点与其相邻数据点的相对紧密程度及类属关系,从而解决密度不均匀聚类问题.依据松散度的性质实现了一种基于松散度的聚类方法,以验证松散度度量的有效性.实验结果表明,使用松散度来度量数据点的聚类密度信息可以有效处理密度不均匀聚类问题.

A metric for measuring density in clustering based on local information

The loose degree which measures density in cluster based on implicit local information was developed to solve the clustering problem for density inhomogeneity.The relation among data points in the same cluster can be revealed by the loose degree,and a clustering algorithm based on loose degree was implemented to verify the effectiveness of the loose degree.Experimental results indicate that the loose degree is a effective density metric for clustering variousdensity datasets.