【摘要】 通过利用点对之间距离的排序信息,给出了聚类评判标准公式,设计出了无参化聚类算法。该算法无需用户指定聚类数,而且聚类质量高于K-M eans算法,较好地解决了聚类数依赖问题。更多还原

【Abstract】 In order to solve the problem of dependence on cluster numbers,the clustering algorithm without(parameters) is designed and the Clustering-Criterion is set up according to the sorting information among pairs of points in the thesis.This algorithm doesn’t need to appoint the clustering number and can also obtain the better result than K-Means.更多还原