INTERNATIONAL JOURNAL OF MULTIDISCIPLINARY SCIENCES AND ENGINEERING, VOL. 2, NO. 8, NOVEMBER 2011 [ISSN: 2045-7057] www.ijmse.org 60 Ontology Construction from Relational Database Abdul Jaleel 1 , Shahid Islam 1 , Asim Rehmat 2 , Amjad Farooq 2 and Adnan Shafiq 2 1 Rachna College of Engineering & Technology, Gujranwala 2 University of Engineering &Technology, Lahore Abstract– This research presents our work to construct Ontologies from relational databases. We have reviewed the existing rules for mapping database schema into ontology schema and consequently updated the mapping rules to accommodate the various missing cases. A real world relational model, the university course registration and fee collection, has been used to demonstrate the mapping of relational data model into ontology concepts. Our example relational model contains every possible construct of the database schema and we have demonstrated the conversion of each construct into ontology using our information system example. Keywords– Semantic Web Services, Ontology, Relational Database and Construction I. INTRODUCTION orld Wide Web (WWW) is a huge information space. Most of the information available over the www comes from the relational database system (RDBS). The data within relational database can be used to construct Ontologies which provides shared common meaning and reusable knowledge about a particular domain. Many approaches [6], [8]-[10] have been proposed to construct Ontologies from RDBS using mapping rules. However these mapping rules don’t take into account all scenarios, found in a RDBS. In this paper, we have reviewed and enhanced the existing rules for mapping the RDBS into Ontology schema. The rest of the paper is organized as follows. Section 2 summarizes the previously related research. Mapping rules are given and explained using an information system example, in section 3. Next section presents the discussion and results. Section 5 concludes our work and presents future directions. II. RELATED WORK Soares [2] reviewed the different approaches of ontology development to identify their suitability for the information systems development process. The study revealed that no standard approach exists for the ontological design and development of information systems. Almost 60% of the engineers don’t use any particular methodology to build the Ontologies. An approach was proposed by Stojanovic [3] to make the database-driven web information space visible and machine process-able. First, relational database schema was revealed using reverse engineering. Then relational data base to object- oriented database conversion rules were applied on the relational schema to generate the corresponding Ontologies under the supervision of expert. Irina [8] presented an approach to migrate data-intensive web pages which are based on relational databases, into the ontology based Semantic Web. Their work uses information from web pages to construct the equivalent ontology. However this approach is not efficient because html forms do not fully represent the semantics of the backend database. The work of [4], [6] and [7] explained the process of ontology creation from relation database using mapping rules. They presented a set of rules for mapping database schema into ontology. However, these works lack some specific database cases which we have dealt in our work. III. CONSTRUCTION OF ONTOLOGY FROM RELATIONAL DATABASE This section presents the revised mapping rules to construct the ontology from relational database. As an example, we take a database schema which stores the student related information. The database example taken here, stores the fee records and course enrollment of students at university level (having semester system). Figure: 1 presents the database schema of this scenario. The given schema contains all types of relationships and constraints that may exist in RDBS. We will use this relational schema for the explanation of mapping rules. We have divided the process of constructing ontology from relational database into the following seven steps. Each step explains the mapping process considering different scenarios that may occur within RDBS. Prerequisite: database schema must be in 3NF at minimum.  In our case the database schema is already in 3NF. Step 1: Identify Primary Keys (P. Key) for all Relations  The attributes having symbol before them (in figure: 1), are P. Keys of our relations. Step 2: Identify Foreign Keys (F. Key) for all Relations  The attributes having symbol before them (in figure: 1), are F. Keys of our relations.  The attributes having symbol before them (in figure: 1), are F. Keys as well as P. Keys or part of P. Keys. Step 3: Formation of Ontology classes Rule 1: this rule is composed of following three cases W