Inproceedings,
Combining Metadata and Document Search with Spreading Activation for Semantic Desktop Search
Proc. of ESWC, page 569-583. Springer, (June 2008)
DOI: 10.1007/978-3-540-68234-9_42
Abstract
The Semantic Desktop is a means to support users in Personal Information
Management (PIM). It provides an excellent test bed for Semantic
Web technology: resources (e. g., persons, projects, messages, documents)
are distributed amongst multiple systems, ontologies are used to
link and annotate them. Finding information is a core element in
PIM. For the end user, the search interface has to be intuitive
to use, natural language queries provide a simple mean to express
requests. State of the art semantic search engines focus on fact
retrieval or on semantic document retrieval. We combine both approaches
to search the Semantic Desktop exploiting all available information.
Our semantic search engine, built on semantic teleporting and spreading
activation, is able to answer natural language queries with facts,
e. g., a specific phone number, and/or relevant documents. We evaluated
our approach on ESWC 2007 data in comparison with Google site search.
