Skip to main content

Table 4 Median (minimum, maximum) number of correctly identified non-zero and zero coefficients, N = 500, 1000 simulated datasets

From: Least absolute shrinkage and selection operator type methods for the identification of serum biomarkers of overweight and obesity: simulation and application

Scenario 1 2 3 4 5
Type Non-Zero Zero Non-Zero Zero Non-Zero Zero Non-Zero Zero Non-Zero Zero
Truth 5 95 5 95 5 95 5 95 20 80
Overweight
 LASSO 1 (0,5) 93 (59,95) 2 (0,5) 92 (54,95) 5 (2,5) 82 (50,95) 3 (2,5) 90 (50,95) 11 (5,16) 75 (46,80)
 Adaptive LASSO 2 (0,5) 81 (51,95) 2 (0,4) 84 (55,95) 5 (3,5) 81 (55,95) 4 (1,5) 81 (54,95) 7 (3,14) 68 (45,80)
 Elastic Net 2 (0,5) 85 (0,95) 3 (0,5) 86 (0,95) 5 (3,5) 78 (16,95) 4 (2,5) 87 (36,95) 14 (6,20) 66.5 (0,80)
 Iterated LASSO 1 (0,5) 93 (66,95) 2 (0,5) 92 (62,95) 5 (2,5) 84 (62,95) 3 (1,5) 91 (60,95) 8 (3,14) 76 (52,80)
 Bootstrap-Enhanced LASSO-75 2 (0,5) 87 (69,95) 1 (0,3) 87 (68,95) 5 (3,5) 84 (62,94) 3 (1,5) 85 (65,95) 8 (4,13) 72 (54,80)
 Weighted Fusion 3 (0,5) 85 (0,95) 5 (0,5) 85 (0,95) 5 (3,5) 89 (34,94) 3 (2,5) 89 (0,94) 20 (6,20) 43 (0,80)
Obese
 LASSO 4 (1,5) 84 (55,95) 3 (1,5) 91 (57,95) 5 (4,5) 77.5 (46,94) 4 (3,5) 85 (49,95) 16 (10,20) 67 (42,80)
 Adaptive LASSO 4 (1,5) 80 (48,95) 3 (1,5) 83 (50,95) 5 (4,5) 81 (52,95) 4 (2,5) 82 (50,95) 11 (6,16) 63 (42,80)
 Elastic Net 5 (1,5) 80 (0,95) 3 (1,5) 87 (0,95) 5 (4,5) 74 (26,95) 4 (3,5) 79 (13,95) 17 (10,20) 55 (0,78)
 Iterated LASSO 4 (1,5) 86 (60,95) 3 (1,5) 91 (64,95) 5 (4,5) 84 (64,95) 4 (2,5) 88 (59,95) 12 (6,17) 75 (59,80)
 Bootstrap-Enhanced LASSO-75 4 (2,5) 85 (64,94) 2 (1,5) 87 (72,95) 5 (4,5) 87 (76,95) 4 (2,5) 89 (79,95) 12 (7,17) 76 (70,80)
 Weighted Fusion 5 (1,5) 62 (0,95) 3 (1,5) 94 (0,95) 5 (4,5) 74 (13,88) 4 (3,5) 77 (30,95) 20 (10,20) 51 (0,80)