Table 3. Used models and tuning parameter values used for the binary outcome data.
| Parameter 1 | Parameter 2 | Parameter 3 | Parameter 4 | |
|---|---|---|---|---|
| Linear logistic regression [57] | - | - | - | - |
| Linear discriminant analysis [58] | - | - | - | - |
| L1-logistic regression† [55, 59–60] | λ1 = 0.00001, 0.0001, 0.001, 0.01, 0.1, 0.5, 1, 10, 20, 30, 50, 70, 100, 500, 1,000 | - | - | - |
| L2-logistic regression† [55, 61–62] | λ2 = 0.00001, 0.0001, 0.001, 0.01, 0.1, 0.5, 1, 10, 20, 30, 50, 70, 100, 500, 1,000 | - | - | - |
| Penalized discriminant analysis† [63] | λ = 0.00001, 0.0001, 0.001, 0.01, 0.1, 0.5, 1, 10, 20, 30, 50, 70, 100, 500, 1,000 | - | - | - |
| Random forest [33] | Ntrees: 1,000 | Npredictors: 2, 3, 4, 5, 6, 7, 9, 11, 13, 16 | Node size = 1 | - |
| Stochastic gradient boosting [34] | Max. Ntrees: 1,000 | Interaction depth: 2, 3, 4, 5, 6, 7, 8, 9 | Bag fraction = 0.5 | ν = 0.01 |
| BART [64–65] | Ntrees: 200 | k = 2.0 | Niter:1,000 | Number of burn-in iterations: 100 |
† Models were tried out both with standardized and unstandardized input data.