Skip to main content

Table 1 Information of the training and test datasets

From: HobPre: accurate prediction of human oral bioavailability for small molecules

Cutoff

Data sets

Molecules

Positive

Negative

F = 50%

Training set

1157

536

621

Test set 1

290

169

121

Test set 2

141

90

51

F = 20%

Training set

1142

859

283

Test set 1

287

214

73

Test set 2

133

128

5