Castells et al [4] designed a retrieval system that exploits
ontology based KBs to improve search over large document
repositories. Semantic search is combined with traditional
keyword based retrieval which tolerates sparseness of KB.
The overall retrieval process consists of following steps. This
system takes as input RDF Data Query Language (RDQL)
query and this is executed against the KB. The output of this
step is list of instance tuples that satisfy the query. For this
execution, ontology processing library, Jena Toolkit is used.
Document Annotation is done using semi-automatic
technique. These annotations are given weights based on TFIDF
algorithm. The documents that are annotated with the
instances returned in previous step are presented to the user