Summary of reviewed adaptive hypertext systems
In summary, the above systems adapt the presentation of, or navigation routes through, information by assessing the cognitive state and intentions of their users. This assessment is based on user input and is variously captured by :
The adaptation that results from this assessment is varied. Navigational adaptation features prominently in the HYPERFLEX system, where it is implemented as a suggestion list. One clear advantage of adaptation in this form is that it is unobstructive and can easily be disregarded if not deemed appropriate. However, this approach is limited to navigational adaptation, in a context where the user is knowledgeable enough to recognise from page titles which pages of information are likely to be relevant to them.
In the context of tutoring systems, users are typically unfamiliar with the material, or with at least parts of it, at the beginning of the interaction. One approach to establishing what the user knows is that taken in KN-AHS and ANATOM-TUTOR. Both of these systems use initial interview to gather information, which is then used to classify users into a number of stereotypes. This technique simplifies system design, but the classification may not be ideal for all users, especially those who are borderline between stereotypes. For this reason a mechanism must exist to allow transitions between stereotypes, particularly as users’ goals and knowledge change over time. KN-AHS uses user selection to complement information gained during an initial interview, while ANATOM-tutor bases updates on the user’s answers to questions.
Authored by Serengul Smith
E-mail to:
serengul1@mdx.ac.uk
School of Computing Science Middlesex University
Revised: September 1998