Table 1.
The most statistically significant models generated by AutoQSAR using the complete dataset.
| Training Set (%) | R 2 | SD | Q2 (R2pred) | RMSE | N | Fingerprint |
|---|---|---|---|---|---|---|
| 70 | 0.5378 | 0.2180 | 0.4937 | 0.2123 | 1 | Radial |
| 75 | 0.5997 | 0.2065 | 0.5284 | 0.1994 | 2 | Dendritic |
| 80 | 0.6304 | 0.2022 | 0.6107 | 0.1817 | 5 | MOLPRINT 2D |
R2: coefficient of determination for the training set; SD: standard deviation; Q2: predictive correlation coefficient for the test set (R2pred); RMSE: root-mean-square error for the test set predictions; N: optimum number of components.