Reasoning with user's preferences in ambient assisted living environments

Understanding the importance of preference management in ambient intelligent environments is key to providing systems that are better prepared to meet users' expectations. Preferences are fundamental in decision making, so it is an essential element in developing systems that guides the choices of the users. These choices can be decided through argument(s) which are known to have various strengths, as one argument can rely on more certain or vital information than the other. The analysis of survey conducted on preferences handling techniques in Artificial Intelligence (AmI), indicates that most of existing techniques lack the ability to handle ambiguity and/or the evolution of preferences over time. Further investigation identified argumentation technique as a feasible solution to complement existing work. Argumentation provides a means to deal with inconsistent knowledge and we explored its potentials to handle conflicting users preferences by applying to it several real world scenarios. The exploration demonstrates the usefulness of argumentation in handling conflicting preferences and inconsistencies, and provides effective ways to manage, reason and represents user's preferences. Using argumentation technique, this research provide a practical implementation of a system to manage conflicting situations, along with a simple interface that aids the flow of preferences from users to the system, so as to provide services that are better aligned with the users' behaviour. This thesis also describes the functionalities of the implemented system, and illustrates the functions by solving some of the complexities in users' preferences in a real smart home. The system detects potential conflict(s), and solves them using a redefined precedence order among some preference criteria. The research further show how the implemented Hybrid System is capable of interacting with external source's data. The system was used to access and filter live data (groceries products) of a UK supermarket chain store, through their application programming interface (API), and advise users on their eating habits, based on their set preference(s).

Level dc:type.qualificationlevel

PhD thesis

Grantor dc:publisher.institution

Middlesex University

Year dc:date.issued

2020

Author dc:creator

• Oguego, C.

Identifier

oai:repository.mdx.ac.uk:89z87

OAI identifier oai:identifier

oai:repository.mdx.ac.uk:89z87