Skip to main content

Table 2 Median (minimum, maximum) number of selected non-zero coefficients from 1000 simulated data sets

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

  N = 1000 N = 500
Scenario 1 2 3 4 5 1 2 3 4 5
True number 5 5 5 5 20 5 5 5 5 20
Overweight
 LASSO 7 (0,38) 5.5 (1,41) 19 (4,61) 9 (2,56) 19 (9,56) 4 (0,37) 5 (0,43) 18 (4,50) 8 (2,49) 16 (5,47)
 Adaptive LASSO 16 (1,45) 14 (1,48) 18 (4,49) 19 (2,51) 21 (4,52) 16 (0,46) 12 (1,43) 18 (3,45) 18 (1,45) 19.5 (3,45)
 Elastic Net 12 (0,100) 10 (1,100) 23 (4,62) 13 (2,65) 28 (10,100) 12 (0,100) 12 (0,100) 22 (4,84) 11 (2,62) 27 (7,100)
 Iterated LASSO 7 (0,41) 5 (1,34) 18 (4,46) 9 (2,43) 16 (5,41) 4 (0,32) 4 (0,35) 16 (2,38) 7 (1,39) 13 (3,37)
 Bootstrap-Enhanced LASSO-75 12 (1,28) 10 (1,26) 18 (7,38) 17 (4,40) 20 (10,40) 10 (0,29) 9 (0,28) 16 (5,37) 13 (3,34) 16 (5,35)
 Weighted Fusion 27 (0,99) 27 (1,100) 23 (10,76) 21 (10,74) 34 (18,100) 13 (0,100) 13 (0,100) 11 (5,66) 9 (4,100) 56 (9,100)
Obese
 LASSO 17 (4,51) 7 (2,59) 26 (8,56) 16 (3,58) 37 (20,59) 15 (2,45) 7 (1,40) 22.5 (5,54) 14 (3,50) 28 (14,52)
 Adaptive LASSO 17 (3,50) 17 (1,47) 18 (5,57) 19 (3,47) 35 (14,57) 19 (1,51) 15 (1,48) 19 (4,48) 17 (2,48) 28 (6,48)
 Elastic Net 21 (5,74) 11 (2,75) 29 (6,76) 19 (4,77) 47 (22,100) 20 (2,100) 11 (1,100) 26 (5,74) 19 (3,87) 42 (14,100)
 Iterated LASSO 16 (4,40) 7 (1,44) 20 (5,40) 13 (3,38) 24 (12,40) 14 (1,39) 6 (1,34) 16 (4,36) 11 (2,40) 17 (7,34)
 Bootstrap-Enhanced LASSO-75 17 (5,37) 14 (2,34) 20 (7,39) 16 (5,34) 22 (12,34) 14.5 (5,36) 11 (1,27) 13 (5,24) 10 (3,20) 16 (8,25)
 Weighted Fusion 12 (5,68) 10 (2,99) 43 (5,61) 36 (3,53) 39 (22,99) 38 (1,100) 4 (1,100) 26 (12,87) 22 (4,70) 47 (14,100)