摘要
文本聚类是聚类的一个重要研究分支,在文本处理领域中有着广泛的应用。在描述聚类特征树与动态索引树的文本聚类方法后,将原动态索引树文本聚类方法中的合并阀值由单一线性依赖关系修改为依赖于聚类节点半径值。实验证明,改进后的算法在聚类结果精确率与聚类时间上都有明显提高。
Text clustering is an important research branch in clustering; it has been used in a wide range of application areas.This report describes the clustering feature tree and the dynamic index tree clustering method.The node＇s threshold in Dynamic Indexing Tree was depended on the single linear relationship,we revised to depend on the cluster node radius.After experimental,the experimental results show that the improved algorithm in clustering and cluster precision time has improved significantly.
出处
《电脑开发与应用》
2010年第9期62-65,共4页
Computer Development & Applications
关键词
动态索引树
阀值
层次聚类
节点半径
dynamic index tree
threshold
hierarchical clustering
node radius
