Table 2.

C-indices and calibration slopes for clinical data example in both scenarios considered using different models

Scenario	Measure	Full	BE 0.05	BE 0.5	BE 0.05 + Ridge	BE 0.5 + Ridge	Lasso	Ridge	Elastic
1	C-index	0.599	0.586	0.596	0.586	0.596	0.600	0.601	0.600
2	C-index	0.601	0.574	0.577	0.574	0.578	0.603	0.607	0.600
1	Calibration slope	0.868	0.927	0.884	0.974	0.960	1.012	1.065	1.050
2	Calibration slope	0.500	0.649	0.583	0.708	0.645	0.861	1.162	0.885

Note: The C-indices and calibration slopes presented are corrected for over-optimism via the 0.632 bootstrap. BE 0.05 + ridge and BE 0.5 + ridge refer to ridge regression applied to the variables selected by BE 0.05 and BE 0.5, respectively. BE, backward elimination.