One of the important tasks of knowledge platforms is to include a large number of different viewpoints in a constantly growing and at the same time heterogeneous information pool and to network them in multidisciplinary knowledge structures. It is very difficult to individually map the knowledge of the community members in these structures. This challenge is being met by using various approaches and methods from the areas of knowledge discovery, knowledge visualization and intuitive interfaces.
In this talk the knowledge management architecture of netzspannung.org and knowledge visualization tools that have been developed for expert Communities will be presented. Special focus will be put on the knowledge card system and the semantic map interface for the contextualisation of content and the discovery of new connections in heterogeneous document pools. In addition, the system combines text analysis methods with a self-organizing neuronal network in order to generate clusters of semantically related content. By contrast to other solutions, the visualization of the clustering results in the semantic map interface also provides an overview of the most important themes as well as enabling discovery of the possible relationships between individual documents. Here, an important development aspect was to make it possible for the user to understand an influence the clustering criteria.
The current system allows administrators of the expert communities to influence the clustering criteria via a map trainer interface and thus to create different knowledge cards. Via the semantic map interface that can be accessed online, normal users can explore the
information pool of netzspannung.org using these knowledge cards. Although this means that possible contexts and connections between different thematic areas can be discovered, the knowledge of individual experts never becomes directly accessible. For this purpose two further models for personalising the knowledge cards have been developed. The first model allows the users to define an information pool selected by themselves and to set down the semantic criteria for grouping the content of the documents interactively. The resulting knowledge cards mirror the explicit know-ledge of individual experts and can be made available to other users as information browsers. The second model investigates how the implicit knowledge of different experts can be transferred. For this purpose, personalised knowledge cards resulting from the user interaction with documents are created and integrated into a common navigation structure. This model is being developed in the project AWAKE and applied in netzspannung.org.
Artists / Authors
- Jasminko Novak, FhG-Institut für Medienkommunikation › Biography
- October 25, 2002
Fraunhofer Institute for Media Communication MARS-Exploratory Media Lab
Schloss Birlinghoven, Sankt Augustin, Germany
Redaktion netzspannung.org, Apr 10, 2003
- collaboration |
- datamining |
- information design |
- databank |
- Graphical User Interface GUI |
- knowledge spaces
Additions to Keyword List