Skip to main content

Table 6 Classification performance on the EVEX resource based on the respective event counts in the EVEX database. This table shows how well the classifier will perform the prediction, preserving correct and eliminating incorrect respective events from the EVEX database. The prediction measures in this table are calculated based on the values in the second column of Table 4. The last column (Support) shows that there are 1,529 incorrect and 4,717 correct corresponding events in the EVEX database (6,246 in total) which are extracted based on those 596 trigger words in the test set

From: Filtering large-scale event collections using a combination of supervised and unsupervised learning for event trigger classification

 

Precision

Recall

F2-score

Support

Negative (incorrect)

0.86

0.44

0.49

1529

Positive (correct)

0.84

0.98

0.95

4717

Weighted averages, total

0.85

0.77

0.77

6246