Skip to main content
. 2018 Sep 19;3(9):11392–11406. doi: 10.1021/acsomega.8b01647

Table 7. Results for BACE1 Dataset (True External Dataset 2) with Best Weighting (0.5–0–0.5) Combination as Obtained from the Retrospective Study.

  training set
 test set
 true external test set
model R2 Q2 MAE95%Train QExt-F12 MAE95%Test correct prediction in % number of compounds outside ADa QExt-F12 MAE95%Test correct prediction in % number of compounds outside ADa
1 0.83 0.76 0.37 0.75 0.32 86.36 0 0.90 0.31 82.35 2 (81, 91)
2 0.80 0.75 0.37 0.79 0.32 91.30 0 0.72 0.37 88.24 2 (81, 89)
3 0.80 0.76 0.35 0.91 0.24 95.65 0 0.83 0.34 88.24 1 (81)
4 0.76 0.71 0.38 0.77 0.27 91.30 0 0.75 0.37 88.24 2 (81, 89)
5 0.79 0.75 0.38 0.79 0.28 91.30 0 0.86 0.33 82.35 1 (81)
a

AD using standardization technique is used and compound ID mentioned under parenthesis; division method: Kennard–Stone; NTraining = 51, NTest = 23, NTrue-External-Test = 17.