Theory and Modern Applications
Table 3 Results of “diabetes” dataset
(ν,N,L) | Training accuracy / Test accuracy / AUC | |
|---|---|---|
Proposed method | (0.01,2,8) | 0.758 / 0.788 / 0.854 |
(0.01,5,8) | 0.758 / 0.788 / 0.854 | |
(0.01,3,24) | 0.823 / 0.801 / 0.899 | |
(0.01,3,80) | 0.838 / 0.913 / 0.968 | |
(0.01,30,8) | 0.758 / 0.788 / 0.854 | |
Existing methods | SVC | 0.778 / 0.789 / 0.746 |
RFC | 0.762 / 0.792 / 0.757 | |
LightGBM | 0.957 / 0.775 / 0.746 | |
XGBoost | 0.857 / 0.784 / 0.753 |