Fuzziness in adaptive hypermedia models

Diverse forms of imperfect information arise in the context of adaptive hypermedia systems, and especially in personalized Web applications, where the user population is fairly diverse. Several adaptive hypermedia systems have been developed that deal explicitly with some form of uncertainty or imprecision, but hypermedia models largely ignore these aspects. In this paper, an extension of the static part of a hypermedia model is described, to address some forms of extended information. Specifically, fuzzy set theory is used to extend some of the model elements, including user categories and links, resulting in more realistic information definitions and also in novel concepts, like fuzzy links, that are generalizations of well-known adaptive techniques. An extension of the UML notation is also described, to visually represent imprecision when modeling an adaptive system.