Research on ontology-based relational data integration

Abstract

To integrate relational data semantically, schema description and query accessing issues are discussed. Ontology is chosen as the global schema to describing semantics. To building domain ontology rapidly, an approach based on subgraph extraction is presented. Then the semantics of SAPRQL queries is defined and algorithm for query rewriting is put forward. Minimal connectable unit is defined to represent a fragment of semantic mapping from the query to the rewriting. Finding and joining minimal connectable units produce semantically correct query plans. Soundness and completeness of query rewriting algorithm is proved. Computational complexity of the algorithm is discussed. Under the worst case, the query rewriting algorithm requires O(n m ) time. Experiment results show that the query processing algorithm can provide nice performance. © 2011 Springer-Verlag.