ANALYSIS AND APPLICATION OF SEMANTIC WEB MECHANISM FOR STORING AND QUERYING ONTOLOGIES
Vaal University of Technology
Since the introduction of Semantic Web in the early 2000, storing and querying ontologies have been a subject of greater research. Thus, several types of storage media and mechanisms have been developed to increase storage and query speed and efficiency of ontologies in Semantic Web applications. Existing Semantic Web mechanisms for storing and querying ontologies are implemented on several storage media and support different languages. However, there is a shortage of studies that provide an empirical analysis and application of these ontology storage and query mechanisms in the Semantic Web domain. This study conducted an analysis and application of the Semantic Web mechanisms for storing and querying ontologies. A thorough literature review was carried out to identify relevant publications pertaining to existing Semantic Web mechanisms for storing and querying ontologies as well as the platforms and storage media for implementing these mechanisms. Thereafter, the Design research method was used consisting of a set of predefined steps, namely, awareness, suggestion, development, evaluation, and conclusion. The awareness stage identified the need for an architecture to test several ontology storage media and mechanisms. In the suggestion stage a framework was proposed to empirically analyse and evaluate existing ontology storage and query mechanisms. The required Semantic Web platforms were identified to implement the framework in the development stage. The evaluation stage used a set of metrics to evaluate the framework including: the loading times of ontologies, the disc space used to store the ontology repositories and the mean and variance of query response times. Further, the evaluation stage analysed and discussed the storage mechanisms implemented in Semantic Web platforms. Finally, the outcome of the performance of the framework is presented in the conclusion stage. The framework was practically tested with six ontologies of different formats and sizes on two popular Semantic Web platforms, namely, Sesame and Jena API and the ontology storage and query mechanisms were analysed and compared. Although the underlying structures of repositories in the in-memory and native files in Jena and Sesame could not be accessed, it was possible to access and analyse the data in the repositories in the relational database storage in both Sesame and Jena. The results showed that Sesame relational uses a combination of mechanisms such as normalized triples store in combination with vertical partitioning. That combination allows vii Sesame to store ontologies based on their contents; in other words, each ontology has a different database schema in Sesame. Jena on the other hand, uses only a normalized triple store mechanism, also known as generic schema mechanism to store ontologies; thus, all ontologies in Jena have the same database schema. The study would be useful to the Semantic Web and Computer Science communities as it does not only provide theoretical knowledge but also the empirical findings that may serve as a base for further development of ontology storage media and mechanisms.