Semantically driven approach for scenario recognition in the IE system FRET

This paper reports a research effort in scenario recognition task in information extraction. The presented approach uses partial semantic analysis based on logical form representation of the templates and the processed text. It is implemented in the system FRET1 (Football Reports Extraction Templates), which processes specific temporally structured texts. A logical inference mechanism is used for filling template forms. Only scenariorelevant relations between events are linked in the inference chains. The knowledge base plays an important role in this process. Some aspects of negation and modalities that occur in the texts are also taken into account.

• Type: Conference
• Year: 2003
• Where: RANLP
• Authors: Svetla Boytcheva, Milena Yankova, Albena Strupchanska
• Keywords: Logical Form Representation, Logical Inference Mechanism, Natural Language Processing, RANLP 2003, Scenario Recognition Task