Skip to main content

Table 2 Statistics of DDI corpus

From: Multiple sampling schemes and deep learning improve active learning performance in drug-drug interaction information retrieval analysis from the literature

Data Source

Sample pool

Data set

Sample size

Initial training set

Initial validation set

PubMed

Screened sample pool

Labeled Positive

150

100 +*

50 +*

Labeled Negative

799

100 -*

50 -*

Unlabeled screened samples

3,169

50 R*

Unscreened sample pool

Unlabeled unscreened samples

9,999

100 +*

50 +*

100 R*

50 -*

50 R*

  1. +* (labeled positive samples), -* (labeled negative samples), R* (random negative samples)
Back to article page

Journal of Biomedical Semantics

ISSN: 2041-1480