Skip to main content

Table 1 Average true discovery rates (%) and average estimation errors over 100 simulations

From: A new insight into underlying disease mechanism through semi-parametric latent differential network model

 

ZP-DDN

RDN

DDN

p

L ∞

L F

L ∞

L F

L ∞

L F

Average true discovery rates

 50

74.0 (13.6)

83.2 (10.9)

75.6 (14.0)

89.1 (11.3)

45.9 (24.7)

27.8 (17.3)

 80

91.4 (16.4)

99.6 (4.3)

95.2 (14.2)

100.0 (0.0)

44.9 (34.6)

51.0 (42.8)

 100

96.3 (14.1)

100.0(0.0)

99.5 (5.2)

100.0 (0.0)

39.3 (40.3)

50.0 (49.1)

 120

78.8 (16.8)

100.0(0.0)

79.0 (18.2)

100.0 (0.0)

23.4 (41.3)

30.0 (46.3)

Average estimation errors in the elementwise L∞ norm

 50

3.26 (0.41)

2.91 (0.33)

3.08 (0.32)

2.59 (0.35)

2.27 (0.12)

2.41 (0.21)

 80

2.06 (0.28)

1.92 (0.06)

1.98 (0.21)

1.91 (0.00)

1.97 (0.09)

1.94 (0.08)

 100

1.86 (0.15)

1.82 (0.00)

1.82 (0.04)

1.82 (0.00)

1.87 (0.10)

1.83 (0.04)

 120

1.12 (0.17)

0.87 (0.00)

1.12 (0.18)

0.87 (0.00)

0.89 (0.07)

0.87 (0.00)