Using Ontology to Improve Precision of Terminology Extraction from Documents

Wen Zhang,Taketoshi Yoshida,Xijin Tang
DOI: https://doi.org/10.1016/j.eswa.2008.12.034
IF: 8.5
2009-01-01
Expert Systems with Applications
Abstract:In this paper, we proposed a new approach using ontology to improve precision of terminology extraction from documents. Firstly, a linguistic method was used to extract the terminological patterns from documents. Then, similarity measures within the framework of ontology were employed to rank the semantic dependency of the noun words in a pattern. Finally, the patterns at a predefined proportion according to their semantic dependencies were retained and regarded as terminologies. Experiments on Retuers-21578 corpus has shown that WordNet ontology, that we adopted for the task of extracting terminologies from English documents, can improve the precision of classical linguistic method on terminology extraction significantly.
What problem does this paper attempt to address?