Personalized Navigation in the Semantic Web Michal Tvaroek
Personalized Navigation in the Semantic Web Michal Tvarožek Mentor: prof. Mária Bieliková Faculty of Informatics and Information Technologies Slovak University of Technology
Motivation l Large open information spaces l Navigation related problems Recursion rate ~ 60% l “Lost in hyperspace” syndrome l Ineffective navigation history l l Search related problems Many irrelevant search results l Difficult construction of semantic queries l
Proposed solution: Enhanced faceted semantic browser l Faceted l Classification Faceted browser l Semantic Web Ontologies l Semantic search l l Adaptive Web Systems Adaptive navigation l Adaptive presentation l
Faceted browser characteristics Faceted navigation l Multi-level facets l Equally weighted facets or Primary and Secondary facets l Search results processing l l Sorting Comparing Views
User interface – concept design
Faceted browser enhancements Ontological database support l Semantic queries (simple/complex, reasoning) l Advanced query mode l l l Combination of multiple facet values Additional combination functions NOT, OR, () Indirect definition of facet values by defining facet value restrictions Built-in support for event logging l User actions, the corresponding server events and their semantics
User adaptation overview l Adaptive navigation Adaptation of facets and facet values l Annotation of search results l l Adaptive presentation of search results Different views for the search results l Highlighting of relevant information l
User interface – concept design
User model l Ontological representation l Derived from a domain ontology l Defines user goals l Defines user characteristics Relation to a specific concept l Relevance to a specific goal l Confidence in its certainty l
Adaptation of facets l Evaluating facet suitability (relevance) Sorting facets and facet values l Selecting relevant items l Emphasizing most relevant items l Hiding irrelevant items l l Dynamic facet generation from the domain ontology and user model l Utilizing concept similarity and collaborative filtering
Annotation of facets, search results l Presentation of additional information Background/text color l Font size l Emoticons, traffic lights l l Evaluation of Gain provided by using a facet l Relative number of instances in facets l Concept similarity (“ideal” search results) l
Prototype implementation – Architecture overview
Prototype implementation – User interface
Summary l Enhanced faceted semantic browser Faceted browsing l Ontologies & semantic search l Adaptive navigation & presentation l l Prototype implementation l Design and implementation of user adaptation l Evaluation in the domain of job offers
- Slides: 14