Applied Research Of Ontology Storage Based On Object-Oriented Database

With the increasing of ontology resources in Semantic Web, the body size of ontology becomes increasingly large and the structure becomes more complex. At this point, the query and management efficiency of ontology becomes a widespread concern. And the storage of ontology directly affects the efficiency of ontology query. Therefore, how to reasonably store large bulk ontology and improve the efficiency of ontology query and search is an interesting and challenging task.Recent years, domestic research on the ontology storage has been continually developing. However, many studies are limited based on relational database, only improve the way of existing storage. The efficiency of ontology query isn't significantly improved.To further improve the efficiency of ontology query, this dissertation explores the ontology storage based on object-oriented database, and we design and implement the storage mode using excellent open source object database db4o and ontology development kit Jena. In Object-oriented database, the data stored directly in the form of object, so can avoid spliting ontology into triples and solve the impedance mismatch problem. More importantly, the ontology query in object-oriented database do not need use the intermediate language such as SQL and intermediate interfaces such as JDBC. So using object database to store ontology is not only more intuitive, but also more efficiency on query compared with the relational database. The main research results of this dissertation are as follows:(1) Discuss the feasibility of ontology storage based on object-oriented database in theory. First we analyze the similarity between object database model and ontology model, and discusse how to use the data manipulation and management capabilities of object database to access the ontology. This work demonstrates the object-based database can greatly improve the efficiency of ontology query.(2) Design and implement a specialized storage using object database db4o and Jena. We use Jena toolkit to create and populate ontology model, and modify its internal interfaces to persistent ontology data to the object database. The system architecture is designed three-layer, and the persistence of ontology data to object database is encapsulated. We only provide interfaces of access ontology object to applications. So the ontology data operation in object database and in memory is nearly the same.(3) Test and analyze the system performance and the efficiency of ontology storage and query. This dissertation analyzes the performance of this storage scheme, tests the efficiency of ontology storage and query under different circumstances and compares the efficiency with relational database.