中文摘要：

提出了一种基于频繁子树模式的GML文档结构聚类算法GCFS（GML Clustering based on Frequent Subtree patterns）,与其他相关算法不同,该算法首先挖掘GML文档集合中的最大与闭合频繁Induced子树,并将其作为聚类特征,根据频繁子树的大小赋予不同的权值,采用余弦函数定义相似度,利用K-Means算法对聚类特征进行聚类。实验结果表明算法GCFS是有效的,具有较高的聚类效率,性能优于其他同类算法。

英文摘要：

This paper presents algorithm GCFS for clustering GML document structure based on frequent subtree patterns.It firstly mines all maximal and closed frequent Induced subtrees from GML documents;then chooses some subtree patterns to form the clustering features,weights these features according to the length of subtree pattern,computes the similarity of two GML documents by cosine function,uses K-Means algorithm to cluster documents by clustering features.Experiment results show that GCFS is effective and efficient.Its performance is superior to other GML clustering algorithms.