Skip to main content

Table 3 Average number and percentage of biologically relevant variables in the model with S N R=0.5 and (β j =0.1,j=1,2,…20)

From: Tilting the lasso by knowledge-based post-processing

Over 100 runs

Adaptive lasso

B1

B2

B3

Average number of selected variables

41

41

41

41

Average number of Biologically relevant variables

12

38

34

29

Average percentage of Biologically relevant variables (%)

29.3 %

92.7 %

82.9 %

70.7 %

Standard deviation

9.1

0.85

1.22

3.08

PMSE (absolute)

1.148

1.145

1.143

1.154

PRPMSE %

100 %

99.7 %

99.6 %

100.5 %

(St.dev)

 

(0.89)

(1.53)

(5.98)

Favorable substitution %

 

91 %

78 %

69 %

(St.dev)

 

(0.99)

(1.07)

(6.93)

MISE

1.572

1.598

1.605

1.643

Over 100 runs

Lasso

B1

B2

B3

Average number of selected variables

53

53

53

53

Average number of Biologically relevant variables

12

46

39

34

Average percentage of Biologically relevant variables (%)

22.6 %

86.8 %

72.1 %

64.2 %

Standard deviation

10.2

0.97

1.53

3.89

PMSE (absolute)

1.576

1.570

1.567

1.596

PRPMSE %

100 %

99.6 %

99.4 %

101.3 %

(St.dev)

 

(0.97)

(1.81)

(6.71)

Favorable substitution %

 

88 %

71 %

63 %

(St.dev)

 

(1.02)

(1.52)

(7.68)

MISE

1.876

1.914

1.927

1.941

  1. Percentage and standard deviations are over 100 runs from data \(\mathcal {D}_{2}\). The average of the PMSE and PRPMSE over 100 runs and the percentage of such runs for which the bootstrap 95 % CI includes 1 or less than 1 and mean integrated squared error (MISE), with S N R=0.5 and (β j =0.1,j=1,2,…,20) from data \(\mathcal {D}_{3}\) using correlation structure for simulation study 1