An adaptable relation extraction system for the biomedical domain is presented. The system makes use of a large set of contextual and shallow syntactic features, which can be automatically optimised for each relation type. The system is tested on three different relation types; protein-protein interactions, tissue expression relations and fragment to parent protein relations.
|Title of host publication||Proceedings of the Workshop on Current Trends in Biomedical Natural Language Processing|
|Place of Publication||Columbus, Ohio|
|Publisher||Association for Computational Linguistics|
|Number of pages||9|
|Publication status||Published - 1 Jun 2008|