Skip to main content

Table 2 Performance of the three other tools (ProClust, TribeMCL and JACOP) and our four proposed methods on DS1, DS2, DS3 and DS4 data sets with respect to two clustering quality measurements: Sensitivity (Sens.) and Specificity (Spec.)

From: Partitioning clustering algorithms for protein sequence data sets

Algorithms

DS1

DS2

DS3

DS4

 

Sens.

Spec.

Sens.

Spec.

Sens.

Spec.

Sens.

Spec.

ProClust

50.64

56.77

48.71

61.86

46.09

55.14

46.39

51.07

TribeMCL

46.09

52.89

41.42

52.14

41.04

47.48

51.22

56.46

JACOP

99.92

66.27

99.96

70.06

99.96

73.96

99.92

94.42

Pro-Kmeans

92.38

99.90

55.32

98.01

63.30

96.92

56.06

99.56

Pro-LEADER

90.21

91.40

53.15

91.24

52.96

74.06

23.34

95.70

Pro-CLARA

93.60

99.92

73.28

99.26

81.53

98.60

77.84

99.66

Pro-CLARANS

93.10

99.90

78. 62

98.70

76.24

97.34

62.06

99.09

  1. DS4 is a very large data set which contains all sequences of DS1 (HLA protein family), DS2 (Hydrolases protein family) and DS3 (Globins protein family).