The objective of this paper is to validate a developed approach to semantic search that builds on a
concept of feature vectors (FV). The approach is based on a pragmatic use of ontologies by relating the
concepts (domain semantics) with the actual terminology used in a text corpus, i.e., the Web. Therefore,
we propose to associate every entity (classes and individuals) of the ontologies with a FV to tailor them
to the domain terminology in a text corpus. First, these FVs are created off-line and later used on-line to
filter, and hence disambiguate search, and re-rank the search results from the underlying search system
(Tomassen & Strasunskas, 2009b).