TOWARDS SEMANTIC WEB - An Effective Approach for Association Identification in Real Scenarios

Abstract

The semantic web is an extension of current web. The concept introduced by this web is dramatically changing the world of web. This is one of the reasons that it is often named as future web. The semantic web’s objective is to make web meaningful and understandable. In the last few years, this area has gained interest from many researchers, bringing the results in the form of different standards for the semantic web. The aim of this web is to make web meaningful, understandable and machine processable. Making use of information on current web for the productiveness of future web is becoming vital. Heterogeneity of relational data coupled with present web complicates utilization of information for future web. To use the data associated with current web it was required to transform it into ontology. The already presented and provided algorithms merely give the result in user required form. The thesis focuses on the problem where some of the meta data is available with relational schema and identification of associations. Protégé is an open source tool that can be used for ontology development. The DataMaster was developed in BioSTORM [10] project which supports both OWL and frame-based ontologies. It works as a plug-in for protégé [11]. Many of these approaches merely give the results in the user required form. Moreover, the weak entities of the database are also mapped into classes (for example in datamaster, DataGenie etc). It will be a tough job to find the relationship in the extracted classes (i.e. ontologies), with a little domain knowledge. Moreover it will become very tough when the relational schema is of large size. Need of an automated or semi automated approach for discovering the relationships in the relational schemas was vital. This presents a scheme for the identification of association in real scenarios, where there can be very little metadata availability.