Skip to main content

Table 3 The frequencies of the relevant variables obtained by the sparse logistic regressions with the L 1/2 , L EN and L 1 penalties in 30 runs

From: Sparse logistic regression with a L1/2 penalty for gene selection in cancer classification

  Sample size Method      
ρ = 0.1 , σ = 0.2 n=50 L1/2 21 22 19 15 15
LEN 24 25 21 17 17
L1 22 24 20 15 17
n=80 L1/2 30 30 30 30 30
LEN 30 29 30 30 30
L1 30 29 30 30 30
n=100 L1/2 30 30 30 30 30
LEN 30 30 30 30 30
L1 30 30 30 30 30
ρ = 0.1 , σ = 0.6 n=50 L1/2 17 17 17 14 14
LEN 18 19 17 16 14
L1 18 18 18 16 15
n=80 L1/2 30 29 30 28 28
LEN 30 28 30 28 27
L1 30 28 30 27 26
n=100 L1/2 30 30 30 30 30
LEN 30 30 30 30 30
L1 30 30 30 28 30
ρ = 0.4 , σ = 0.2 n=50 L1/2 19 18 18 16 15
LEN 21 22 21 17 17
L1 18 21 19 16 17
n=80 L1/2 30 30 30 30 30
LEN 30 28 30 29 29
L1 30 27 30 29 29
n=100 L1/2 30 30 30 30 30
LEN 30 30 30 30 30
L1 30 30 30 29 29
ρ = 0.4 , σ = 0.6 n=50 L1/2 14 16 15 12 12
LEN 17 17 17 12 14
L1 17 15 14 9 13
n=80 L1/2 29 25 26 28 29
LEN 28 24 24 27 24
L1 27 24 24 23 23
n=100 L1/2 30 29 30 30 30
LEN 30 27 28 28 30
   L1 29 27 27 28 30