Skip to main content

Table 5 Results of the complete process for CS and MATH target concepts using as evaluation metrics P (precision), TP (true positives), and FP (false positive)

From: Exploring knowledge graphs for the identification of concept prerequisites

  

CS

MATH

  

FP

TP

P

FP

TP

P

CM

SemRefDHW>0.1

90

147

0.62

54

175

0.76

 

SemRefDHW>0.2

54

85

0.61

44

101

0.70

 

SemRefDHW>0.3

6

11

0.65

5

14

0.74

 

SemRefDNHW>0.1

31

153

0.83

23

168

0.88

 

SemRefDNHW>0.2

13

80

0.86

10

87

0.90

 

SemRefDNHW>0.3

3

35

0.92

2

41

0.95

 

SemRefDJW>0.1

65

149

0.70

44

166

0.79

 

SemRefDJW>0.2

27

84

0.76

22

100

0.82

 

SemRefDJW>0.3

11

33

0.75

8

37

0.82

LC (lmax=1)

SemRefDHW>0.1

67

110

0.62

54

129

0.70

 

SemRefDHW>0.2

16

38

0.70

14

43

0.75

 

SemRefDHW>0.3

6

15

0.71

4

17

0.81

 

SemRefDNHW>0.1

19

129

0.87

17

150

0.90

 

SemRefDNHW>0.2

10

72

0.88

10

87

0.90

 

SemRefDNHW>0.3

3

37

0.93

2

44

0.96

 

SemRefDJW>0.1

40

122

0.75

25

148

0.86

 

SemRefDJW>0.2

22

67

0.75

15

75

0.83

 

SemRefDJW>0.3

8

26

0.76

6

28

0.82

LC (lmax=2)

SemRefDHW>0.1

104

144

0.58

78

172

0.69

 

SemRefDHW>0.2

56

95

0.63

42

102

0.71

 

SemRefDHW>0.3

14

25

0.64

11

28

0.72

 

SemRefDNHW>0.1

34

199

0.85

26

210

0.89

 

SemRefDNHW>0.2

18

97

0.84

15

108

0.88

 

SemRefDNHW>0.3

3

39

0.93

2

44

0.96

 

SemRefDJW>0.1

82

192

0.70

53

224

0.81

 

SemRefDJW>0.2

38

119

0.76

23

144

0.86

 

SemRefDJW>0.3

9

36

0.80

6

38

0.86

  1. Best values per metric and search/pruning strategy are bold