Skip to main content

Table 5 Effectiveness of clustering of MDDR dataset using QPI: ECFP_4 Fingerprint

From: Voting-based consensus clustering for combining multiple clusterings of chemical structures

Clustering method

No. of clusters

500

600

700

800

900

1000

Consensus clustering

CVAA

Correlation

74.86

78.02

82.39

84.16

85.71

87.04

Cosine

74.79

78.12

81.85

84.78

85.91

87.18

Euclidean

71.04

74.92

78.41

81.91

84.47

86.80

Hamming

70.99

74.36

78.47

81.68

84.24

86.28

Jaccard

83.48

87.01

88.72

90.98

90.67

92.05

Manhattan

70.74

74.26

78.52

81.74

84.12

86.09

CSPA

Correlation

70.58

73.29

74.86

76.86

79.17

82.03

Cosine

71.23

71.85

76.43

76.55

78.06

81.21

Euclidean

65.33

67.09

72.49

72.73

74.50

78.75

Hamming

64.68

66.82

69.88

71.25

74.17

76.64

Jaccard

69.91

71.73

74.20

76.01

77.72

79.26

Manhattan

63.07

65.77

68.83

71.50

74.06

77.33

HGPA

Correlation

72.61

74.85

76.4

78.32

80.22

82.26

Cosine

72.06

74.25

77.21

79.54

81.02

83.31

Euclidean

70.71

72.82

75.02

76.80

80.50

82.66

Hamming

69.45

72.21

74.08

77.71

79.67

82.36

Jaccard

67.88

70.58

73.93

76.56

77.65

79.67

Manhattan

72.74

72.14

75.68

77.94

81.42

82.97

Individual clustering

Ward's method

 

75.83

79.88

83.34

84.25

86.49

88.25