Eliminating Incorrect Events from Large‐Scale Event Networks by Trigger Word Clustering and Pruning
: Farrokh Mehryary, Suwisa Kaewphan, Kai Hakala, Filip Ginter
: Bodenreider Olivier, Oliveira José Luis, Rinaldi Fabio
: International symposium on semantic mining in biomedicine
: 2014
: Proceedings of the 6th International Symposium on Semantic Mining in Biomedicine (SMBM 2014)
: 75
: 79
DOI: https://doi.org/10.5167/uzh-98982
In this short paper, we investigate hierarchical
clustering of event triggers in the EVEX large-scale event resource. As the primary application, we utilize the clustering to identify incorrect trigger event words and subsequently eliminate events extracted with these triggers. We evaluate the method on the BioNLP 2011 and 2013 Shared Task test sets and show that the method can further increase the precision and F-score of the winning system of the 2013 BioNLP Shared Task on Event extraction.
