Todays web embraces enormous amount of data. To search for valid, updated and trustworthy information is being more and more complicated. Semantic web brings some promising prospects to this almost hopeless situation.
Various techniques has been developed to annotate data scaterred across the Internet. Such annotations or metadata allow us express semantic of the web pages or exactly define terms and relationships within a specific domain. However creating such an extensive description brings annother level of complexity. To create, use and simply to understand these descriptions is not a trivial task.
The primary aim of this project was to invent, evaluate, implement and deploy a visualization technique which would be able to address the problem of enormity and complexity of formal specifications used to describe the content of web pages.
Very comprehensive analysis of nuts and bolts of visualization of large information spaces has been provided. The thesis describes most used and most popular tools and approaches, consequently provides a design of simple, yet effective visualization technique.
Incremental Browsing with Metadata Importance Evaluation
Incremental browsing of an information space is based on visualization of limited section of the whole graph. The rest of the graph is visualized when needed. The limited section of the information space is called a window and the process of browsing means the movement of the window along a chosen trajectory.
Of course, not all elements of the window are diplayed all the time, to decide when to display these elements we have to estimate relative importance of each element. Importance estimation is not a tivial task. We provide simplified classification of possible evaluation methods. A combination of these methods has been implemented into the prototype.
SemVis - Prototype of Inceremtntal Ontology Browser
In order to experiment with the proposed method we developed software prototype of ontology browser named SemVis. We experimented in the domain of job offers by taking advantage of the domain ontology created in the NAZOU research project aimed at retrieving information in a heterogeneous information space.
SemVis is basically a web application which provides textual and graphical way to browse selected ontologies. Elements of browsed ontology are evaluated to estimate their importance and allow to adapt content of the actual window.
Proposed ontology visualization method brings two significant advantages: