Table 8.
Mean prediction error (MPE) and root mean square error (RMSE) for the final stepwise model and the single-objective, hybrid genetic algorithm (SOHGA) candidate model with lowest fitness function value
| Compound | Mean prediction error (MPE) | Root mean square error (RMSE) | ||||
|---|---|---|---|---|---|---|
| Final stepwise model | Best SOHGA candidate model | abs(MPESOHGA) − abs(MPEstepwise) | Final stepwise model | Best SOHGA candidate model | RMSESOHGA − RMSEstepwise | |
| Citalopram | 0.46 | 0.28 | −0.18 | 92.2 | 53.3 | −38.9 |
| DMAG | −5.49 | −4.51 | −0.98 | 1,167.8 | 1,195.1 | 27.3 |
| Escitalopram | −1.16 | −1.06 | −0.10 | 132.6 | 133.1 | 0.5 |
| Olanzapine | 0.52 | −1.26 | 0.74 | 345.4 | 367.1 | 21.7 |
| Perphenazine | −0.19 | −0.17 | −0.02 | 14.9 | 15.1 | 0.2 |
| Risperidone | −0.34 | 0.81 | 0.47 | 116.9 | 116.5 | −0.4 |
| Ziprasidone | 8.14 | 8.31 | 0.17 | 622.0 | 613.8 | −8.2 |