Paper: A Hybrid approach for biomedical event extraction

ACL ID W13-2017
Title A Hybrid approach for biomedical event extraction
Venue Proceedings of the BioNLP Shared Task 2013 Workshop
Session
Year 2013
Authors

In this paper we propose a system which uses hybrid methods that combine both rule-based and machine learning (ML)-based approaches to solve GENIA Event Extraction of BioNLP Shared Task 2013. We apply UIMA1 Frame- work to support coding. There are three main stages in model: Pre-processing, trigger detec- tion and biomedical event detection. We use dictionary and support vector machine classifi- er to detect event triggers. Event detection is applied on syntactic patterns which are com- bined with features extracted for classification.