Skip to main content

Table 5 RF model performance measures based on DTC-split model 1 and all-split model 5

From: Nonadditivity in public and inhouse data: implications for drug design

ChEMBL data Train R2 (RMSE) Test R2 (RMSE) Test MCC
A* NA# A* NA#
1613777 DTC-split 0.91 (0.17) 0.56 (0.44) − 0.43 (1.30) 0.48 0.02
All-split 0.64 (0.47) 0.22 (0.68) − 0.34 (1.25) 0.34 0.00
1613797 DTC-split 0.66 (0.22) 0.05 (0.41) − 0.29 (1.14) 0.45 − 0.03
All-split 0.43 (0.45) 0.05 (0.58) − 0.31 (1.11) 0.07 0.00
  1. Bold values are best performance measures across DTC-split and All-split
  2. Train R2 is based on 5-fold cross validation results
  3. *Additive test data
  4. #Nonadditive test data