Paper: Improving Feature-Based Biomedical Event Extraction System by Integrating Argument Information

ACL ID W13-2015
Title Improving Feature-Based Biomedical Event Extraction System by Integrating Argument Information
Venue Proceedings of the BioNLP Shared Task 2013 Workshop
Session
Year 2013
Authors

We describe a system for extracting biomedi- cal events among genes and proteins from biomedical literature, using the corpus from the BioNLP?13 Shared Task on Event Extrac- tion. The proposed system is characterized by a wide array of features based on dependency parse graphs and additional argument informa- tion in the second trigger detection. Based on the Uturku system which is the best one in the BioNLP?09 Shared Task, we improve the per- formance of biomedical event extraction by reducing illegal events and false positives in the second trigger detection and the second ar- gument detection. On the development set of BioNLP?13, the system achieves an F-score of 50.96% on the primary task. On the test set of BioNLP?13, it achieves an F-score of 47.56% on the primary task ob...