Skip to main content
. 2010 Sep 4;26(18):i554–i560. doi: 10.1093/bioinformatics/btq381

Fig. 1.

Fig. 1.

Overview of the general event extraction pipeline. For each event type, candidate events in the training data are used to create a feature selector, which is subsequently applied for FS of both training and testing instances. Finally, a classifier is built with the filtered training samples and applied for predicting events in the test set.