A semantic case based web 2.0 tag hierarchy construction framework

In contrast to tags with no inter-relation, web users are more familiar with concept systems with hierarchies, which could also be leveraged in various information retrieval applications. The construction of tag hierarchies relies on the analysis of tag semantic characteristics, while such characteristics are evidenced by semantic case objects which are co-related to tags. Based on such an observation, a semantic case based tag hierarchy construction framework is proposed, together with a semantic flow analysis based hierarchy construction algorithm. Experiment results show that the proposed framework provides an effective mechanism for studying and comparing available algorithms, and the semantic flow based hierarchy construction algorithm outperforms the other evaluated methods in various evaluation metrics.