In recent years, knowledge extraction from texts is
focused to overcome to bottleneck of building ontologies for
semantic web. Pattern based, template driven and linguistic
methods are among the successful approaches to extract
ontological knowledge from raw texts. This paper introduces
some lexico-syntactic and semantic patterns and templates for
extracting conceptual knowledge from Persian texts. The
described patterns are general and domain/ application
independent and work at sentence level. They are used to extract
taxonomic and non-taxonomic relations and axioms from phrases
and sentences. Among the introduced patterns and templates,
semantic patterns are language independent and although
linguistic (lexico-syntactic) patterns are introduced for Persian
language, they could easily be adopted to other languages such as
English.
This paper will first have a brief overview on linguistic and
template driven methods to discover ontological knowledge from
texts. Then the templates for Persian simple sentences will be
introduced. The extracted relations may be hyponymy,
meronymy, attribute/value, part-of, equivalency, etc. The
introduced templates exploit semantic information to not only
extract the instances of predefined relations (the concepts related
by them), but also define new relations. In each case some
examples of the experimental results will make the patterns clear.
