An Approach for Deep Web Interface Schema Extraction Based on Hierarchical Semantic Annotation

摘要

Interface schema extraction plays an important part in the process of Deep Web data source integrating. Referencing query interfaces' hierarchical mapping characteristics while analyzing the semantic relationships among attributes, with an RIT(Relation Indexed Tree) model constructed up , this paper proposed a method using hierarchical semantic annotation to supervise over the attributes' pre-clustering with the precision greatly enhanced. Effectivity of the invented approach has also been proved according to the experiments over several domains in ICQ dataset.