Open Access
Table 5
Performances of SVR and XGBoost-based QSPR models when applied to SDB2 Test sets. The last column reports R2 values when QSPR models are used to predict compounds involved in mixtures of the SDB1 dataset.
| Model | Parameters | R2 Test | R2 SDB1 |
|---|---|---|---|
| Moreno’s SVR [3] | cost = 537.930000 | 0.891 | 0.949 |
| ϵ = 2.305940 | |||
| γ = 0.419818 | |||
| SVR | cost = 45.515217 | 0.950 | 0.962 |
| ϵ = 0.004023 | |||
| γ = 0.000202 | |||
| XGBoost | n estimators = 290 | 0.967 | 0.971 |
| max dept = 3 | |||
| learning rate = 0.278645 | |||
| subsample = 0.906254 |
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.