Learning Object Metadata Mapping for Linked Open Data

Database is a great part of data store on the Internet. However, the transition to web of data and shareable contents to open standard like LOD requires new structuring of traditional data. In order to achieve this interoperability of information systems, RDF is important file format in resolving semantic heterogeneity. This paper proposes the metadata mapping processes to convert database to RDF. DB2RDF is a tool that automatically generates RDF from database by mapping from LO ontology and stored in TDB. The mapping process start by extract database schema and restructuring in form of RDF using Jena API make them available to query. The created model modification query uses SPARQL queries. We proceed set of class and properties from database attributes and redefine with specific domain control vocabulary to populate RDF. Finally, we test the query results that can preserved the database semantic by validated the resulting by SPARQL queries with RDB on SQL queries.