Skip to main content

Table 6 Machine learning experiment test results on the data representation models of the full gene BOUN10CANCER dataset

From: Statistical representation models for mutation information within genomic data

Algorithm

Data Rep.

Accuracy

F-Score

Precision

Recall

Roc-Auc

FPR

NB

binary

33.84 ± 0.83

35.25 ± 0.95

37.04 ± 1.34

33.84 ± 0.83

0.62 ± 0.02

8.38 ± 0.11

 

c-score

31.10 ± 0.86

32.72 ± 0.74

34.53 ± 1.43

31.10 ± 0.86

0.59 ± 0.01

8.61 ± 0.08

 

tf-idf

33.34 ± 0.48

35.04 ± 0.60

37.03 ± 1.03

33.34 ± 0.48

0.62 ± 0.02

7.99 ± 0.07

 

tf-rf

38.14 ± 0.57

38.97 ± 0.87

40.08 ± 1.27

38.14 ± 0.57

0.65 ± 0.01

7.99 ± 0.10

 

bm25-tf-idf

32.50 ± 0.96

34.19 ± 0.87

36.08 ± 1.35

32.50 ± 0.96

0.60 ± 0.01

8.48 ± 0.10

 

bm25-tf-rf

37.94 ± 0.63

38.99 ± 0.60

40.12 ± 1.24

37.94 ± 0.63

0.62 ± 0.01

7.91 ± 0.10

KNN

binary

11.54 ± 0.85

16.87 ± 0.66

31.46 ± 2.54

11.54 ± 0.85

0.50 ± 0.04

7.41 ± 0.04

 

c-score

15.87 ± 0.63

22.60 ± 0.44

39.27 ± 4.21

15.87 ± 0.63

0.53 ± 0.01

7.96 ± 0.07

 

tf-idf

34.96 ± 0.66

37.35 ± 0.95

38.92 ± 0.69

34.96 ± 0.66

0.62 ± 0.03

8.10 ± 0.04

 

tf-rf

19.29 ± 0.44

22.23 ± 0.61

40.29 ± 0.82

19.29 ± 0.44

0.55 ± 0.02

7.57 ± 0.07

 

bm25-tf-idf

12.72 ± 1.23

20.05 ± 0.58

47.32 ± 5.85

12.72 ± 1.23

0.51 ± 0.01

8.17 ± 0.37

 

bm25-tf-rf

11.91 ± 1.13

19.21 ± 0.50

49.74 ± 1.58

11.91 ± 1.13

0.51 ± 0.01

7.88 ± 0.17

SVM-poly

binary

17.50 ± 0.00

5.21 ± 0.00

3.06 ± 0.00

17.50 ± 0.00

0.53 ± 0.00

16.34 ± 0.00

 

c-score

56.14 ± 0.44

58.90 ± 0.39

61.96 ± 0.46

56.14 ± 0.44

0.73 ± 0.01

5.33 ± 0.06

 

tf-idf

17.50 ± 0.00

5.21 ± 0.00

3.06 ± 0.00

17.50 ± 0.00

0.53 ± 0.00

16.35 ± 0.00

 

tf-rf

55.51 ± 0.55

56.52 ± 0.65

61.40 ± 0.53

55.51 ± 0.55

0.71 ± 0.03

5.16 ± 0.05

 

bm25-tf-idf

36.36 ± 0.66

42.64 ± 0.75

51.56 ± 0.89

36.36 ± 0.66

0.62 ± 0.01

7.93 ± 0.08

 

bm25-tf-rf

53.41 ± 0.27

51.46 ± 0.27

63.95 ± 0.65

53.41 ± 0.27

0.66 ± 0.01

7.38 ± 0.04

SVM-rbf

binary

66.71 ± 0.36

67.01 ± 0.00

68.01 ± 0.00

67.01 ± 0.01

0.78 ± 0.01

4.04 ± 0.09

 

c-score

57.35 ± 0.30

61.31 ± 0.28

65.86 ± 1.10

57.35 ± 0.30

0.72 ± 0.01

7.09 ± 0.05

 

tf-idf

50.92 ± 0.19

44.26 ± 0.20

51.64 ± 0.19

50.92 ± 0.19

0.69 ± 0.02

8.30 ± 0.03

 

tf-rf

69.53 ± 0.71

69.82 ± 0.72

70.75 ± 0.71

69.53 ± 0.71

0.78 ± 0.03

3.64 ± 0.09

 

bm25-tf-idf

66.17 ± 0.56

66.61 ± 0.60

67.20 ± 0.62

66.17 ± 0.56

0.78 ± 0.01

4.40 ± 0.07

 

bm25-tf-rf

73.77 ± 0.46

74.00 ± 0.46

74.96 ± 0.40

73.77 ± 0.46

0.83 ± 0.01

3.20 ± 0.07

SVM-linear

binary

68.46 ± 0.67

68.01 ± 0.01

69.01 ± 0.01

68.01 ± 0.01

0.78 ± 0.01

4.07 ± 0.09

 

c-score

71.91 ± 0.44

72.46 ± 0.45

73.02 ± 0.44

71.91 ± 0.44

0.82 ± 0.01

3.50 ± 0.09

 

tf-idf

69.54 ± 0.66

69.01 ± 0.01

70.01 ± 0.01

69.01 ± 0.01

0.78 ± 0.01

3.94 ± 0.06

 

tf-rf

68.80 ± 0.62

68.01 ± 0.01

69.51 ± 0.01

69.01 ± 0.01

0.78 ± 0.01

3.74 ± 0.09

 

bm25-tf-idf

66.26 ± 0.58

66.35 ± 0.60

67.94 ± 0.66

66.26 ± 0.58

0.78 ± 0.01

4.31 ± 0.07

 

bm25-tf-rf

73.44 ± 0.43

73.66 ± 0.45

74.63 ± 0.41

73.44 ± 0.43

0.83 ± 0.01

3.24 ± 0.07

LR

binary

67.19 ± 0.41

68.01 ± 0.01

68.01 ± 0.00

67.01 ± 0.01

0.78 ± 0.01

3.85 ± 0.07

 

c-score

73.50 ± 0.64

73.89 ± 0.92

74.29 ± 0.66

73.50 ± 0.64

0.83 ± 0.01

3.40 ± 0.08

 

tf-idf

63.17 ± 0.30

60.01 ± 0.00

66.01 ± 0.01

63.01 ± 0.00

0.74 ± 0.01

5.68 ± 0.04

 

tf-rf

71.51 ± 0.46

72.01 ± 0.01

73.01 ± 0.01

71.01 ± 0.01

0.81 ± 0.01

3.24 ± 0.07

 

bm25-tf-idf

67.80 ± 0.45

68.20 ± 0.47

68.61 ± 0.53

67.80 ± 0.45

0.79 ± 0.01

4.09 ± 0.06

 

bm25-tf-rf

74.99 ± 0.41

75.19 ± 0.38

75.96 ± 0.37

74.99 ± 0.41

0.83 ± 0.01

3.03 ± 0.06

Perceptron

binary

68.50 ± 0.48

69.01 ± 0.01

70.01 ± 0.01

68.01 ± 0.01

0.78 ± 0.03

4.07 ± 0.09

 

c-score

71.64 ± 1.54

71.76 ± 1.87

71.89 ± 1.38

71.64 ± 1.54

0.81 ± 0.01

3.67 ± 0.24

 

tf-idf

70.23 ± 0.40

70.01 ± 0.00

70.01 ± 0.01

70.01 ± 0.01

0.79 ± 0.01

3.83 ± 0.05

 

tf-rf

72.07 ± 1.86

72.01 ± 0.02

74.01 ± 0.01

72.01 ± 0.02

0.82 ± 0.02

3.29 ± 0.12

 

bm25-tf-idf

65.52 ± 0.52

65.97 ± 0.52

66.44 ± 0.56

65.52 ± 0.52

0.78 ± 0.01

4.48 ± 0.08

 

bm25-tf-rf

74.15 ± 0.51

74.48 ± 0.56

75.46 ± 0.56

74.15 ± 0.51

0.83 ± 0.01

3.07 ± 0.10

Feed-Forward NN

binary

69.00 ± 0.76

69.52 ± 0.70

71.00 ± 0.52

69.00 ± 0.81

0.79 ± 0.02

3.65 ± 0.17

 

c-score

73.74 ± 0.88

74.07 ± 0.73

74.41 ± 0.67

73.74 ± 0.88

0.84 ± 0.02

3.27 ± 0.24

 

tf-idf

62.91 ± 0.79

63.32 ± 0.70

65.04 ± 0.52

62.91 ± 0.83

0.73 ± 0.02

4.00 ± 0.10

 

tf-rf

74.13 ± 1.33

74.17 ± 1.47

75.43 ± 1.07

74.13 ± 1.40

0.85 ± 0.02

3.07 ± 0.24

 

bm25-tf-idf

68.18 ± 1.83

68.79 ± 1.28

69.42 ± 0.76

68.18 ± 1.83

0.82 ± 0.02

4.07 ± 0.54

 

bm25-tf-rf

76.44 ± 0.66

76.95 ± 0.68

77.48 ± 0.78

76.44 ± 0.66

0.86 ± 0.02

2.75 ± 0.13

  1. The row with the best accuracy and f-score is shown in italic for each algorithm. The overall best performance is made bold