Skip to main content

Table 7 Median delays F-scores of case III using long time series with D-CLINDE

From: High-order dynamic Bayesian Network learning with hidden common causes for causal gene regulatory network

n

n h

p bias

T

Complete (C)

Hidden (H)

IgnoreHidden

H/C

p-value

50

5

0.65

100

0.526

0.259

0.339

49.2 %

—

   

200

0.723

0.435

0.510

60.2 %

—

   

400

0.841

0.590

0.611

70.1 %

—

   

800

0.898

0.757

0.647

84.3 %

6.37E-12

   

1000

0.906

0.777

0.660

85.8 %

9.09E-13

   

1200

0.911

0.806

0.662

88.5 %

9.09E-13

   

1400

0.916

0.822

0.660

89.7 %

9.09E-13

   

1600

0.923

0.839

0.660

90.9 %

9.09E-13

  

0.75

100

0.669

0.356

0.455

53.1 %

—

   

200

0.812

0.488

0.579

60.1 %

—

   

400

0.864

0.676

0.631

78.3 %

2.20E-05

   

800

0.905

0.782

0.643

86.5 %

9.09E-13

   

1000

0.911

0.818

0.636

89.9 %

9.09E-13

   

1200

0.910

0.828

0.629

91.0 %

1.85E-08

   

1400

0.913

0.831

0.635

91.1 %

9.09E-13

   

1600

0.917

0.828

0.634

90.3 %

9.09E-13

  

0.85

100

0.725

0.422

0.520

58.2 %

—

   

200

0.823

0.554

0.597

67.4 %

—

   

400

0.884

0.702

0.634

79.5 %

1.74E-08

   

800

0.911

0.803

0.641

88.1 %

9.09E-13

   

1000

0.915

0.796

0.638

87.0 %

9.09E-13

   

1200

0.915

0.825

0.629

90.2 %

9.09E-13

   

1400

0.917

0.813

0.629

88.7 %

9.09E-13

   

1600

0.912

0.818

0.625

89.7 %

9.09E-13

100

10

0.65

100

0.494

0.290

0.310

58.6 %

—

   

200

0.708

0.387

0.494

54.6 %

—

   

400

0.824

0.571

0.600

69.3 %

—

   

800

0.883

0.715

0.639

81.0 %

2.73E-12

   

1000

0.896

0.758

0.642

84.5 %

9.09E-13

   

1200

0.900

0.790

0.649

87.8 %

9.09E-13

   

1400

0.911

0.796

0.652

87.4 %

9.09E-13

   

1600

0.913

0.801

0.646

87.8 %

9.09E-13

  

0.75

100

0.647

0.362

0.442

56.0 %

—

   

200

0.795

0.515

0.577

64.8 %

—

   

400

0.864

0.692

0.630

80.1 %

8.22E-10

   

800

0.900

0.784

0.633

87.1 %

9.09E-13

   

1000

0.909

0.794

0.634

87.4 %

9.09E-13

   

1200

0.916

0.805

0.638

87.9 %

9.09E-13

   

1400

0.917

0.817

0.635

89.2 %

9.09E-13

   

1600

0.921

0.827

0.632

89.9 %

9.09E-13

  

0.85

100

0.705

0.419

0.505

59.4 %

—

   

200

0.813

0.542

0.582

66.7 %

—

   

400

0.883

0.690

0.624

78.1 %

1.82E-12

   

800

0.912

0.766

0.627

84.1 %

9.09E-13

   

1000

0.917

0.784

0.622

85.5 %

9.09E-13

   

1200

0.919

0.778

0.622

84.7 %

9.09E-13

   

1400

0.920

0.798

0.618

86.7 %

9.09E-13

   

1600

0.926

0.790

0.616

85.3 %

9.09E-13

  1. Complete is D-CLINDE on the complete data. hidden is our proposed algorithm with D-CLINDE on the incomplete data. ignoreHidden is D-CLINDE on the incomplete data. p-value is for one-sided Wilcoxon signed rank test on whether the median F-score of hidden is better than ignoreHidden, and entries larger than 0.1 are omitted. H/C is the ratio of hidden over complete as percentage.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us