Abstract: The solution to semantic heterogeneity of the integrated data is a key point in the integration of heterogeneous databases. To aim at adding semantic information to tables and fields in relational databases,automatic semantic annotation for metadata in relational database is a hot spot. In this paper, a semantic annotation method was developed based on similarity algorithms. Firstly,implicit semantic information is extracted from metadata in relational databases,and domain ontologies are created based on the information. Then, the extracted information is annotated automatically by using name similarities and structural similarities between metadata and the ontology entity, and the similarity algorithms are improved. The experiment shows that the method is effective for the task of automatic semantic annotation.

Key words: Semantic annotation, Metadata, Ontology, Database