Skip to main content

Table 2 Mean (SD) of final model size for the training sets over 300 simulations

From: Combined performance of screening and variable selection methods in ultra-high dimensional data in predicting time-to-event outcomes

Selection approach Weak signal Strong signal
n = 150 n = 300 n = 150 n = 300
ISIS-LASSO 1.087 (1.130) 6.010 (0.672) 2.567 (2.265) 6.073 (0.261)
ISIS-ALASSO 0.993 (0.925) 5.943 (0.732) 2.397 (2.164) 6.033 (0.180)
ISIS-RSF 2.463 (0.550) 6.007 (0.954) 3.367 (1.777) 6.033 (0.304)
SIS 2.277 (0.448) 3.120 (0.925) 2.477 (0.500) 4.380 (0.955)
SIS-LASSO 0.773 (0.955) 3.033 (1.014) 1.217 (0.945) 4.377 (0.951)
SIS-ALASSO 0.750 (0.737) 3.003 (1.049) 1.100 (0.824) 4.373 (0.951)
SIS-RSF 2.057 (0.644) 3.033 (1.037) 2.240 (0.719) 4.327 (0.961)
PSIS 132.183 (11.557) 120.440 (10.707) 129.980 (11.203) 119.103 (10.702)
PSIS-LASSO 69.850 (7.238) 78.503 (7.301) 63.170 (7.212) 65.473 (6.786)
PSIS-ALASSO 52.613 (6.344) 49.007 (7.881) 42.487 (8.784) 32.067 (9.986)
PSIS-RSF 44.26 (10.512) 16.033 (5.713) 42.667 (11.009) 15.113 (5.159)
LASSO 26.619 (11.320) 61.437 (13.310) 58.347 (10.184) 108.470 (14.364)
ALASSO 19.948 (7.318) 45.010 (8.276) 33.793 (7.738) 67.147 (9.422)
  1. SD standard deviation