Ontology-based semantic heterogeneous data integration framework for learning environment

• Main Author: Yunianta, Arda
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/54898/1/ArdaYuniantaPFC2015.pdf; http://eprints.utm.my/id/eprint/54898/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:95170

Nowadays, e-learning has become important supporting tools for effective learning. Therefore, integrating a good learning environment in e-learning can improve learning process. Good learning environment can provide new knowledge. Currently, there are many distributed systems and applications on learning environment that involve heterogeneity data in data level implementation. Different learning applications have different system designs and data representations. The main problem on learning environment is that every individual learning application has limited capability to share data and information. Moreover, existing data integration approaches still have weaknesses and there has been less research done on the learning environment of data integration. This research proposes a semantic data integration framework is to handle data heterogeneity on learning environment that integrates various learning information to produce new learning knowledge. This research focuses on semantic data integration using an ontology approach to handle semantic relationship between data sources. The research methodology consists of three main stages. The first stage is semantic data mapping to standardize the heterogeneity data representation from numerous data sources into a standard file format. The second stage is to design and develop the ontology knowledge to create semantic relationship between different data sources. The third stage is to combine the ontology knowledge with the semantic data mapping file to produce semantic data integration. Ontology validation process on this framework uses Resource Description Framework (RDF) validator by World Wide Web Consortium (W3C) standardization and Factplusplus (FaCT++) reasoning in order to check the consistency of classes, instances and properties. Moreover, to validate the framework, this research employs the quality criteria and the metric, based on the Quality Framework for Data Integration approach. The quality criteria focus on the completeness and consistency of the data sources, while the metric produces the quality factor to determine the degree of acceptance. This framework is then verified by adding three different learning systems with heterogeneity in data level implementation which are the Moodle e-learning system, the Question Bank system and the Student Grading system. This framework successfully integrates the different data sources with heterogeneity data representation using quality factor formulas and the result shows that this framework is capable to produce new learning knowledge that involves complex learning information.