Paper: Identification of Genia Events using Multiple Classifiers

ACL ID W13-2018
Title Identification of Genia Events using Multiple Classifiers
Venue Proceedings of the BioNLP Shared Task 2013 Workshop
Session
Year 2013
Authors

We describe our system to extract genia events that was developed for the BioNLP 2013 Shared Task. Our system uses a su- pervised information extraction platform based on Support Vector Machines (SVM) and separates the process of event clas- sification into multiple stages. For each event type the SVM parameters are ad- justed and feature selection carried out. We find that this optimisation improves the performance of our approach. Overall our system achieved the highest precision score of all systems and was ranked 6th of 10 participating systems on F-measure (strict matching).