【摘要】 针对K-means算法所存在的问题进行了深入的研究,提出了基于密度和聚类对象方向的改进算法(KADD 算法)．该算法采取聚类对象分布密度方法来确定初始聚类中心,然后根据对象的聚类方向来发现任意形状的簇．理论分析与实验结果表明,改进算法在不改变时间、空间复杂度的情况下能取得更好的聚类结果．更多还原

【Abstract】 The existing problems of K-means clustering algorithm were carefully researched.An improved K-means algorithm based on density and direction(KADD) was presented,with which the initial clustering center points were located according to the clustering objects distribution density.And the clusters with arbitrary distributions were found based on object direction.Theory analysis and experimental results demonstrate that the improved algorithm can get better clustering without changing efficiency and dimensional complexity.更多还原