Biomedical events describe complex interactions between various biomedicalentities. Event trigger is a word or a phrase which typically signifies theoccurrence of an event. Event trigger identification is an important first stepin all event extraction methods. However many of the current approaches eitherrely on complex hand-crafted features or consider features only within awindow. In this paper we propose a method that takes the advantage of recurrentneural network (RNN) to extract higher level features present across thesentence. Thus hidden state representation of RNN along with word and entitytype embedding as features avoid relying on the complex hand-crafted featuresgenerated using various NLP toolkits. Our experiments have shown to achievestate-of-art F1-score on Multi Level Event Extraction (MLEE) corpus. We havealso performed category-wise analysis of the result and discussed theimportance of various features in trigger identification task.
展开▼
