Skip to main content

Table 2 Effectivenss of clustering of high diverse MDDR dataset: ECFP_4 Fingerprint.

From: Using graph-based consensus clustering for combining K-means clustering of heterogeneous chemical structures

Clustering Method

No. of clusters

   

500

600

700

800

900

1000

Consensus (CSPA)

 

N = 5

57.36

61.39

65.25

68.93

71.90

74.69

  

N = 10

58.51

64.01

67.98

70.23

75.04

75.79

  

N = 15

61.28

64.45

68.16

71.27

73.44

74.34

  

N = 20

60.78

64.92

68.70

71.22

74.37

74.45

  

N = 25

62.03

65.88

68.46

71.11

75.04

74.27

  

N = 30

61.85

64.64

67.27

70.17

73.35

76.01

  

N = 35

62.23

65.91

68.44

71.30

72.97

73.75

  

N = 40

61.67

64.62

67.79

69.31

73.61

74.92

  

N = 45

61.80

65.11

67.96

71.37

74.07

75.41

  

N = 50

60.91

64.96

68.56

70.57

74.57

73.33

Individual

Wards' Method

 

64.86

68.89

74.12

76.09

79.13

82.23