Semantic querying and search in distributed ontologies

Aslam, Sharjeel (2021) Semantic querying and search in distributed ontologies. Doctoral thesis, London Metropolitan University.

Abstract

We have observed in recent years a continuous growth in the quantity of RDF data accessible on the web. This evolution is primarily based on increasing data on the web by different sectors such as governments, life science researchers, or academic institutes. RDF data creation is mainly developed by replacing existing data resources with RDF, changing relational databases into RDF. These RDF data are usually called qualified linked data URIs and endpoints of SPARQL. Continuous development that we are experiencing in SPARQL endpoints requires accessing sets of distributed RDF data repositories is getting popularity. This research has offered an extensive analysis of accessing RDF data across distributed ontologies. The existing approaches lack a broad mix of RDF indexing and retrieving of distributed RDF data in one package. In addition, the efficiency of the current methods is not so dynamic and mainly depend on manual fixed strategies for accessing RDF data from a distributed environment. The literature review has acknowledged the need for a robust, reliable, dynamic, and comprehensive accessing mechanism for distributed RDF data using RDF indexing. This thesis presents the conceptual framework that demonstrates the SPARQL query execution process, which accesses the data within distributed RDF sets across a stored index. This thesis introduces the semantic algebra involved in the conversion of traditional SPARQL query language into different phases. The proposed framework elaborates the concepts included in selecting, projection, joins, specialisation and generalisation operators. These operators are usually in assistance during the process of processing and converting a SPARQL query. This thesis introduces the algorithms behind the proposed conceptual framework, which covert the main SPARQL query into sub-queries, sending each subquery to the required distributed repository to fetch the data and merging the sub queries results. 4 This research demonstrates the testing of the proposed framework using the unit and functional testing strategies. The author developed and utilised the Museum ontology to test and evaluate the developed system. It demonstrates all how the complete developed and processed system works. Different tests have been performed in this thesis, like the algebraic operator's test (e.g., select, join, outer join, generalisation, and specialisation operators test) and test the proposed algorithm. After comprehensive testing, it shows that all developed system units worked as expected, and no errors found during the testing of all phases of the tested framework. Finally, the thesis presents implemented framework's performance and accuracy by comparing it to other similar systems. Evaluation of the implemented system demonstrated that the proposed framework could handle distributed SPARQL queries very effectively. The author selected FedX, ANAPSID and ADERIS existing frameworks to compare with developed system and described the results in a graphical format to illustrate the performance and accuracy of all systems.

Documents
7082:36846
[img]
Preview
Aslam-Sharjeel_Full-thesis.pdf - Published Version

Download (4MB) | Preview
Details
Record
Statistics

Downloads

Downloads per month over past year



Downloads each year

View Item View Item