Skip to main content

Table 1 The scale of positive samples in the warm-start and cold-start setting

From: HDN-DDI: a novel framework for predicting drug-drug interactions using hierarchical molecular graphs and enhanced dual-view representation learning

Setting

Warm-start setting

Cold-start setting

Dataset

DrugBank

Twosides

DrugBank Fold 0

DrugBank Fold 1

DrugBank Fold 2

Training set

115,085

2,745,772

108,396

114,617

110,094

Validation set

38,361

915,257

15,388

11,733

15,217

Testing set

38,362

915,258

68,024

65,458

66,497

   - S1 partition

–

–

7,369

6,719

6,935

   - S2 partition

–

–

60,655

58,739

59,562

Total

191,808

4,576,287

191,808

191,808

191,808

  1. The symbol “–” indicates that there were no S1 and S2 partitions in the warm-start setting
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us