Research And Implementation Of Semantics About Relational Database Based On Ontology

The database is a collection of data, and the database system is an important part ofthe production and living in modern society. So far, relational database systems are themost widely used, and its data model is relational model. The relational model hassimple data structure--two-dimensional table, mature mathematical theory--relationalalgebra, and a series of integrity constraints. Those all make the relational model hasunique advantages in data management and make the relational database system whichbased on it become the norm. So far, thousands of relational database systems have beenestablished all over the world, which have stored the sheer amount of data.However, because of this simple and single data structure, the relational model losta large amount of semantic information during describing fact. Relational modelexpresses the entity and the relationships among entities only by two-dimensional table.It expresses relations only by a subset of Cartesian product. The description of factmade by the data structure as two-dimensional table in relational database is excessiverigid and lack of semantics, which makes a lot of hidden information existingin relational database systems.Aiming at the lack of semantics in relational database system, this paper referenceda classical knowledge representation form—ontology, brought part of its ideas intorelational database. Proposed a semantic model based on ontology, this model addedsemantic layer and reasoning layer on the top of relational data. It implanted part of theideas of ontology into semantic layer and then constructed the semantic meta-structure,which is used to represent the fact. On this basis, we can make reasoning through somerules. This semantic model makes the data in relational database can expresses the factsemantically, and be reasoned correctly and effectively. At the same time, we can getlots of information hidden in relational data after reasoning in the semantic model. Itincreased the integrity of information, improved the availability of relational data, andextended the applicability of RDBMS.