AD vs CN classification on ADNI 20%-PCA based variant (train test)
88.3AUCUG-GEPSVM
Evaluation results
| Method | Links | |
|---|---|---|
| UG-GEPSVMHyperparameters=(δ, hops, γ) = (10−3, 2, 102), Patterns × Features=1017×1552026.06 | 88.3 | |
| UGEPSVMHyperparameters=(δ, γ) = (10−3, 102), Patterns × Features=1017×1552026.06 | 88.2 | |
| IUGEPSVMHyperparameters=(δ, γ, ψ, σ) = (10−1, 10−1, 10−3, 101), Patterns × Features=1017×1552026.06 | 86.68 | |
| IUG-GEPSVMHyperparameters=(δ, γ1, γ2, ψ1, ψ2, hops) = (10−3, 10−3, 10−3, 10−3, 10−3, 4), Patterns × Features=1017×1552026.06 | 86.42 | |
| IGEPSVMHyperparameters=(δ, ν, γ) = (10−3, 10−2, 10−2), Patterns × Features=1017×1552026.06 | 83.1 | |
| GEPSVMHyperparameters=(δ, γ) = (10−3, 10−1), Patterns × Features=1017×1552026.06 | 73.2 | |
| UTSVMHyperparameters=(c, cu, ϵ, reg, γ) = (10−1, 100, 10−1, 100, 10−1), Patterns × Features=1017×1552026.06 | 72.7 |