摘要
BIRCH算法是针对大规模数据集的聚类算法。文章针对该算法在聚类精度和效率方面存在的不足,在原有算法的基础上进行了改进,提出了可变多阈值聚类特征树方法,实现了基于该方法的聚类分析器,最后将该算法应用于图像分割技术当中,验证算法的有效性。
BIRCH algorithm is a clustering algorithm for very large datasets.Considering the imperfections of clustering precision and clustering efficiency which exist in the BIRCH algorithm,this paper presents an ameliorated solution,clus-tering the data objects with convertibility mult_threshold CF-tree,and realizes the clustering analysis.Finally,the algo-rithm is used in image segmentation and proved efficiently.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第12期174-176,195,共4页
Computer Engineering and Applications
