Skip to main content

Table 1 Summary of simulation scenarios

From: Data integration by multi-tuning parameter elastic net regression

Independent Features: ρ1 = ρ2 = ρ12 = 0

Scenario #

β 1

β 2

q 1

q 2

Optimal penalty ratio (κ*)

 1

0.6

0.8

5

20

0.55

 2

0.6

0.6

5

20

0.70

 3

0.8

0.6

5

20

0.75

 4

0.8

0.8

5

5

1

 5

0.8

0.6

5

5

1

Correlated Features: β1 = 0.8, β2 = 0.6, q1 = 5, q2 = 20, r1 = r2 = 3

 

ρ 1

ρ 2

ρ 12

Optimal penalty ratio (κ)

 6

0.4

0.2

0

0.85

 7

0.4

0.2

0.4

0.9

  1. Two hundred samples per data set, 250 features per omic type, 2 omic types. The performance of MTP EN is evaluated by varying effect sizes, number of informative features, and correlation structures between omic types. Specifically, ρ1 is the correlation between informative features in platform 1, ρ2 is the correlation between informative features in platform 2, ρ12 is the correlation between informative features from the different platforms, β1 and β2 are the effect sizes of informative features in platforms 1 and 2, respectively, while q1 and q2 are the numbers of informative features