中文摘要：

提出一种合理运用大众标注数据自动构建资源导航体系的方法，首先使用聚类技术将标签分成主题相近的簇，然后在簇内按标签的泛指度和相似性构造标签树。、将该方法用在公开的标签数据集Social一ODP一2k9上进行实效评价，结果表明，本方法减少了以往标签树构造方法中“意义漂移”的缺陷，使组织结构有更好的语义一致性，且具有更好的导航效率，从而使本研究能适应信息组织与服务的实际需要。

英文摘要：

This paper proposes a new approach to automatically construct the resource navigation system with taxonomy from social annotation. Tags are first clustered to groups according to topics through clustering technology; then a sub - tree is constructed according to tags＇similarity and their generality degree in each group. The open Social - ODP - 2k9 tags dataset are experimented on to make eft~ctiveness evaluation. The method overcomes the shortcoming of semantic -drift in the existing method, and improves the semantic consistency of intbrmation organization and efficiency of navigation, to meet the new practical need for information organization and services.