Personalized Navigation in the Semantic Web An Enhanced

Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser Michal Tvarožek 30. 3. 2006 tvarozek@fiit. stuba. sk, Personalized Navigation in the Semantic Web: An Enhanced FIIT STU BA

Presentation outline • Motivation • Proposed solution • Progress to date • Work to do next • Summary 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 2

Motivation • The navigation problem – “Lost in hyperspace” • Simplify navigation • Large open information spaces • User-friendly semantic query construction • User data acquisition & logging 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 3

Proposed solution – Factic Faceted Semantic Browser • Ontologies and semantic markup – Domain ontology [NAZOU] – User ontology [NAZOU] • Enhanced Faceted Browser – Faceted classification – Multi-level facets – Equally-weighted facets – Adaptivity 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 4

Implementation environment • Java 1. 5 (v Eclipse ) • Apache Cocoon framework • Works as a java servlet in a servlet container (Jetty, Apache Tomcat) • Apache AXIS as a web service “container” • Sesame for ontological storage • Subversion for source code versioning 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 5

Faceted Semantic Browser – Factic • Construction and execution of semantic queries • Browsing and simple processing of search results • Input – Domain ontology (data and faceted classification) – User ontology (in the future) – Configuration files • Output – Interface description (XML data file) 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 6

Semantic Logging – Semantic. Log • Logs events with known semantics • Requires an ontology of events – Ontology in very early design stage • Input: event descriptions (+event ontology) • Output: log file in XML format • Web service interface 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 7

Cooperating user modeling tools • Factic needs a “full” user model • User modeling tools are needed – Semantic. Log supplies server-side log data – Click [Michal Barla] supplies client-side log data – Log Analyzer [Michal Barla] processes acquired data • User model stored in ontological repository 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 8

Progress to date • Basic implementation of a faceted browser – 3 defined facets – Simple semantic queries – Browsing of search results – Simple sorting of search results • Basic implementation of a logging tool – Present implementation must be used directly as a java class – Output is not a full XML file, but an XML fragment • Basic design of a user interface 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 9

How does/will it all work? 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 10

Example of a user interface 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 11

Work to do next • Generalize hard-coded functionality • Generalize facet definitions • Define more static facets, facet types – Enumerations, numeric intervals, dates • Modify Semantic. Log to work as a web service • Adaptation • Identify / define evaluation methods / metrics 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 12

Sources of adaptation • User’s knowledge (what I know) • User’s goals (what I want to do) • User’s background (what I am) • User’s experience (my knowledge of hyperspace) • User’s preferences (what I like) èGoals, background and knowledge are most suitable èPreferences and experience are harder to use 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 13

Adaptation options • Adaptation of predefined facets – Reordering, hiding, annotation of facets and values – Direct or local guidance • Dynamic facet generation – Generation of facets from the domain & user ontology • Adaptation of search results – Annotation, sorting, hiding – Need for external evaluation tools (Concept. Comparer) 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 14

Adaptation of predefined facets • Reordering of facets – Based on attribute relevance • Hiding of facets, facet values – Based on attribute relevance and number/relevance of values • Annotation of facets and facet values – Based on user goals • Recommending restrictions – Based on the greatest reduction of the information space 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 15

Summary • Basic implementation exists • Near-term goals are defined and work is in progress • Adaptation methods must be explored in more detail • Future work – Advanced query mode – History tracking / evaluation – Integration with Prescott [NAZOU] – Integration with searching tools [NAZOU] 30. 3. 2006 Personalized Navigation in the Semantic Web: An Enhanced Faceted Browser 16