Skip to main content

Table 3 Pairwise comparison of imputation method.

From: Comparative analysis of missing value imputation methods to improve clustering and interpretation of microarray experiments

(a)

          
 

kNN

BPCA

Row Mean

EM_gene

EM_array

LSI_gene

LSI_array

LSI_combined

LSI_adaptative

SkNN

kNN

-----

23.47

47.65

60.82

4.59

38.06

5.00

5.41

7.25

47.14

BPCA

-----

-----

75.41

81.33

11.12

67.04

12.76

14.49

16.63

75.51

Row Mean

-----

-----

-----

64.69

4.49

40.82

5.10

5.71

6.12

52.45

EM_gene

-----

-----

-----

-----

3.67

29.49

4.08

4.39

5.31

37.04

EM_array

-----

-----

-----

-----

-----

92.45

60.04

63.89

63.36

95.00

LSI_gene

-----

-----

-----

-----

-----

-----

7.86

7.65

7.45

61.53

LSI_array

-----

-----

-----

-----

-----

-----

-----

37.24

38.27

94.79

LSI_combined

-----

-----

-----

-----

-----

-----

-----

-----

44.39

93.78

LSI_adaptative

-----

-----

-----

-----

-----

-----

-----

-----

-----

92.96

SkNN

-----

-----

-----

-----

-----

-----

-----

-----

-----

-----

(b)

          
 

kNN

BPCA

Row Mean

EM_gene

EM_array

LSI_gene

LSI_array

LSI_combined

LSI_adaptative

SkNN

kNN

-----

42.59

44.02

55.90

6.32

45.45

18.74

18.74

18.74

50.09

BPCA

-----

-----

52.02

63.49

7.84

53.04

23.37

23.37

23.37

58.03

Row Mean

-----

-----

-----

62.18

6.69

24.88

14.01

14.01

14.01

56.58

EM_gene

-----

-----

-----

-----

5.06

39.27

15.67

15.67

15.67

44.54

EM_array

-----

-----

-----

-----

-----

92.97

79.65

79.65

79.65

93.61

LSI_gene

-----

-----

-----

-----

-----

-----

14.85

14.85

14.85

55.52

LSI_array

-----

-----

-----

-----

-----

-----

-----

39.24

43.29

81.67

LSI_combined

-----

-----

-----

-----

-----

-----

-----

-----

46.49

81.67

LSI_adaptative

-----

-----

-----

-----

-----

-----

-----

-----

-----

81.67

SkNN

-----

-----

-----

-----

-----

-----

-----

-----

-----

-----

  1. Is given the percentage of better approximation of one method versus another for a rate of missing value t equal to (a) 32% and (b) 48.5% with the OS dataset. The percentage is given in regards to the method given at the left.