Abstract
The regular indexing of text documents is based on the textual representation and does not evaluate the actual document content. In the semantic web approach, human-authored text documents are transformed into machine-readable content data which can be used to create semantic relations among documents. In this paper, we present on-going work in the WIKINGER project which aims to build a web-based system for semantic indexing of text documents by evaluating manual and semi-automatic annotations. A particular feature is the continuous refinement of the automatically generated semantic network by considering community feedback. The feasibility of the approach will be validated in a pilot application.
