Skip to main content
. 2018 Mar 26;34(17):2973–2981. doi: 10.1093/bioinformatics/bty190

Table 3.

The F1-score when training and testing on the same dataset (source ⇒ source). Specifically, these are the results of training on the 80% split and testing on the 20% split

AIMed BioInfer DDI
CNN 0.5335 0.6079 0.7392
Bi-LSTM 0.6152 0.6562 0.7804