摘要
针对经典的基于密度的聚类分析算法DBSCAN存在的聚类质量对参数敏感,不能处理多密度数据集等问题,提出基于相对密度的聚类分析算法RDCMD,该算法以某点密度与其领域密度的比值表示该点的相对密度,由于不同点的领域信息是不同的,所以相对密度是动态变化的,从而可以适应多密度数据集中点的密度变化。因此,RDCMD算法可以处理多密度数据集,同时能够自动识别噪声。
According to the abuse and problem in traditional Density-based Spatial Clustering of Appli- cations with Noise (DBSCAN), such as clustering quality is sensitive to parameters; multi-den- sity data set cannot be processed, etc., gives Relative Density-based Cluster analysis algorithm (RDCMD). This algorithm uses the ratio of the density at a given point and field density to in- dicate the relative density of that point, due to different point has different field information, and the relative density has the dynamic variation which can adapt to the density change in multi-density data sets. Therefore, RDCMD algorithm can deal with multidimensional data sets, and automatically identify noise as well.
出处
《现代计算机》
2013年第9期3-7,20,共6页
Modern Computer