Integration of the Thesaurus for the Social Sciences (TheSoz) in an Information Extraction System
Thierry Declerck · 2013
We present current work dealing with the integration of a multilingual thesaurus for social sciences in a NLP framework for supporting Knowledge-Driven Information Extraction in the field of social sciences. We describe the various steps that lead to a running IE system: lexicalization of the labels of the thesaurus and semi-automatic generation of domain specific IE grammars, with their subsequent implementation in a finite state engine. Finally, we outline the actual field of application of the IE system: analysis of social media for recognition of relevant topics in the context of elections. 1