Context-informed adaptive hypermedia

Abstract

Adaptive Hypermedia Systems have mechanisms for extensively modelling their domain of expertise, and reasoning over that domain to produce tailored web content. In general, these systems are supplied in advance with deep, explicit models of the relevant axes of adaptivity. Current designs in Adaptive Hypermedia therefore attempt to model the entire sphere of experience of the system deeply and without separation. Contextual input, on the other hand, is arbitrarily defined, and deep models are not separated from `shallow interest? contextual data. It is desirable to model this data separately so as to reduce the size of AH models and to encapsulate low-level sensor data as high-level interpreted concepts. This paper presents a novel method for applying contextual information to an Adaptive Hypermedia eLearning system, providing a clear boundary between `core? adaptive axes and contextual information. An outline of the architecture of an integrated Context Interpretation system with an Adaptive Hypermedia System is provided, along with analysis of the mechanisms by which contextual information may be applied to the specific case of eLearning, and in more general systems.