Extensions to the Model

The book "Dynamic taxonomies and faceted search" presents Chapter 5 of the book "Dynamic taxonomies and faceted search" presents extensions to the dynamic taxonomy model, which include: • data-mining functionalities in dynamic taxonomy exploration, which support both a sophisticated visual environment for data-mining applications, and wisdom-seeking exploration tasks; • the extension to non-atomic, structured objects (such as video material) for which access at different levels of granularity must be supported; • the virtualization of parts of the taxonomy, which are dynamically computed rather than explicitly defined. These extensions (virtual concepts, logics, and web ontologies) increasingly extend the base model from the expressive point of view, by dealing with an infinite number of concepts (such as those which represent numeric or date values), to concepts that are logical formulas, and to taxonomy shapes and extensions defined through logical inference; • fuzzy dynamic taxonomies, in which the membership of object o to concept C is represented by a probability. Fuzzy dynamic taxonomies allow to model several practical situations (e.g., the frequency of a specific symptom in a pathology for medical diagnosis applications) in a more precise way, while retaining the ease of use and the normal operations supported by DTs. In addition, a number of useful techniques are described, which use some of the features of DTs in a non-obvious way. These include: • using predefined foci in order to personalize or control access; • embedding SAES facets in dynamic taxonomies. We consider tag clouds, geographic maps and object clusters; • accounting for popularity, recommendations, and authoritativeness by specific facets or taxonomic ordering; • using a DT as a thesaurus to expand the IR index for a textual object in order to improve IR recall.