Table 3:
Data split for automatic classification: The table is in order of completion of annotation batches.
Annotation batch | Total Articles | Training Articles | Testing Articles |
---|---|---|---|
Prior work: E.K.W., M.R.B.*, (E.D.**) | 52 | 37 | 15 |
Training: Gold-standard, K.J.S., S.A. | 8 | 6 | 2 |
K.J.S., S.A., (M.R.B***) | 8 | 6 | 2 |
K.J.S., M.R.B.*** | 11 | 8 | 3 |
S.A., M.R.B.*** | 12 | 8 | 4 |
Total Articles | 91 | 65 | 26 |
Total Sentences | 12,055 | 8,281 | 3,774 |
Total Words | 416,866 | 285,439 | 131,427 |
Note that E.K.W. is Elizabeth K. White, M.R.B. is Mayla R. Boguslav, E.D. is Emily Dunn, Gold-standard is the previous gold-standard up to that point (the first row), K.J.S. is Katherine J. Sullivan, and S.A. is Stephanie Araki.
M.R.B. is an annotator along with the others.
E.D. only annotated one article along with the other annotators and then stopped.
M.R.B. was the adjudicator in these batches.