Abstract
We are now living in a world where the adoption of pervasive technologies is becoming more prevalent. Personalisation has become a fundamental element for context-aware applications, which has coincided with the increase in the use of mobile-based services and technology dependence. Work presented in this Thesis utilises novel methods that aid the provisioning of context-aware services to suit the dynamicity of people. A review of context-aware technologies highlighted issues involving the modelling of user dynamicity, enabling personalisation based on human contexts and the complexity of knowledge engineering for user modelling. These issues guided the focus of four studies in this Thesis.A study assessing the ability for models to represent users and analyse context-aware environments has been undertaken. As a result, an ontological model was developed to represent user dynamicity and was subsequently used throughout the Thesis as a core component in the methods by which personalisation can be provisioned. In one study, the underlying ontology was evaluated in the context of a ‘Help-on-Demand’ personalisation service and results positively highlighted the use of semantic rules to provide efficient delivery of personalised media to users.
In a subsequent study, an automation tool that uses technology to mine and subsequently learn user behaviours was developed to enable the creation of ontology models. The study established that based on mined data, profiles could be successfully output in a semi-automated manner. The tool was evaluated with comprehensive datasets to aid non-expert users in maintaining their own user profiles.
The final study aimed to overcome the issue of traditional complex knowledge engineering, through the creation of a visual programming interface for ontologies. This tool was evaluated with non-expert users and both learnability and functionality where assessed, with results indicating a positive impact of the tool on all levels of users.
The incorporation of these findings with the development of components for contextaware services will help to ensure that underlying user models can be designed to facilitate ‘on-demand’ user personalisation.
| Date of Award | Feb 2016 |
|---|---|
| Original language | English |
| Supervisor | Luke Chen (Supervisor), Christopher Nugent (Supervisor) & Mark Donnelly (Supervisor) |
Cite this
- Standard