MLPQNA performance against training set size variation. As features we used the best 32 found by the ΦLAB method and as target the given SFRs. The statistics is calculated on the blind test set.
| Number of training objects . | RMSE . | Median . | η . |
|---|---|---|---|
| . | . | . | . |
| 36 000 | 0.337 | −0.015 | 1.53 |
| 100 000 | 0.281 | −0.017 | 1.62 |
| 362 208 | 0.248 | −0.017 | 1.99 |
| Number of training objects . | RMSE . | Median . | η . |
|---|---|---|---|
| . | . | . | . |
| 36 000 | 0.337 | −0.015 | 1.53 |
| 100 000 | 0.281 | −0.017 | 1.62 |
| 362 208 | 0.248 | −0.017 | 1.99 |
MLPQNA performance against training set size variation. As features we used the best 32 found by the ΦLAB method and as target the given SFRs. The statistics is calculated on the blind test set.
| Number of training objects . | RMSE . | Median . | η . |
|---|---|---|---|
| . | . | . | . |
| 36 000 | 0.337 | −0.015 | 1.53 |
| 100 000 | 0.281 | −0.017 | 1.62 |
| 362 208 | 0.248 | −0.017 | 1.99 |
| Number of training objects . | RMSE . | Median . | η . |
|---|---|---|---|
| . | . | . | . |
| 36 000 | 0.337 | −0.015 | 1.53 |
| 100 000 | 0.281 | −0.017 | 1.62 |
| 362 208 | 0.248 | −0.017 | 1.99 |
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
