Table 5.
Classification with the CAESAR model.
| Training set | Observed logBCF | First validation set | Observed logBCF | Second validation set | Observed logBCF | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 327 comp. | nB | B | vB | 81 comp. | nB | B | vB | 119 comp. | nB | B | vB | |||
| Predicted logBCF | nB | 82.46 (270) |
3.38 (11) |
0.31 (1) |
Predicted logBCF | nB | 90.00 (72) |
3.75 (3) |
0.00 (0) |
Predicted logBCF | nB | 88.24 (105) |
4.20 (5) |
0.84 (1) |
| B | 1.54 (5) |
2.15 (7) |
0.92 (3) |
B | 0.00 (0) |
1.25 (2) |
1.25 (1) |
B | 0.84 (1) |
1.68 (2) |
2.52 (3) |
|||
| vB | 0.62 (2) |
1.23 (4) |
7.38 (24) |
vB | 1.25 (1) |
0.00 (0) |
2.50 (2) |
vB | 0.00 (0) |
0.84 (1) |
0.84 (1) |
|||
Three sets are reported: training, first validation and second validation set. The percentage of the total of compounds predicted is given without considering those outside the applicability domain. In brackets, the number of compounds for each class. The total number of compounds is also reported.