Context-Aware Modeling Using Semantic Web and Z Notation

Surveys in user context modeling have shown that the semantic web is one of the promising approach to represent and structure the contextual information captured from user's surrounding environment in a context-aware application. A benefit of using semantic web language is that it enables application to reason user contextual information in order to get the knowledge of user's behavior. However, regarding its notation format, semantic web is suitable for implementation level or to be consumed by application run-time. Context-aware application is a part of distributed computing system. In distributed computing system, the language used for specification should be distinguished from the implementation I run-time purpose. This is known as separation of modeling language. Regarding the context-aware application, for those who are concerned with specification of context modeling, the language that is used for specification should also be distinguished from the implementation one. This thesis aims at proposing the use of formal specification technique to develop a generic context ontology model of user's behavior at the Computer and Information Sciences Department, Universiti Teknologi PETRONAS. Initially, the context ontology was written in OWL semantic web language. The further process is mapping onto a formal specification language, i.e. onto Z notation. As a result, specification of context ontology and its consistency checking have been developed and verified beyond the semantic web language environment. An inconsistency of context model has been detected during the verification of Z model, which cannot be revealed by current OWL DL reasoner. The context-aware designers might benefit from the formal specification of context ontology, where the designers could fully use formal verification technique to check the correctness of context ontology. Thus, the modeling approach in this thesis has shown that it could complement the context ontology development process, where the checking and refinement are performed beyond the semantic web reasoner