Skip to main content

Table 14 SMO classifiers’ results

From: "Our Little Secret": pinpointing potential predators

 

Filter

Stopwords filter

Attributes Selection

TP

FP

FN

Pre.

Recall

F1

SVM, Unigram

1.

No Filter

No

No – 703

110

16

32

0.87

0.77

0.82

2.

No Filter

No

Yes – 569

106

16

36

0.87

0.75

0.80

3.

No Filter

Yes

No – 764

99

15

43

0.86

0.70

0.77

4.

No Filter

Yes

Yes – 572

98

17

44

0.85

0.70

0.76

5.

Computer

No

No – 654

113

11

29

0.91

0.80

0.85

6.

Computer

No

Yes – 576

114

10

28

0.92

0.80

0.86

7.

Computer

Yes

No – 654

106

12

36

0.90

0.75

0.82

8.

Computer

Yes

Yes – 586

106

12

36

0.90

0.75

0.82

SVM, Trigram

9.

No Filter

No

No – 743

104

17

38

0.86

0.73

0.79

10.

No Filter

No

Yes – 639

104

16

38

0.87

0.73

0.79

11.

No Filter

Yes

No – 783

106

13

36

0.89

0.75

0.81

12.

No Filter

Yes

Yes – 648

105

13

37

0.89

0.74

0.81

13.

Computer

No

No – 720

107

12

35

0.90

0.75

0.82

14.

Computer

No

Yes – 602

104

13

38

0.89

0.73

0.80

15.

Computer

Yes

No – 758

102

11

40

0.90

0.72

0.80

16.

Computer

Yes

Yes – 648

105

13

38

0.89

0.74

0.81

SVM Applied on Accept file of our Approach

17.

Computer

NA

No – 136

61

10

81

0.86

0.43

0.57

18.

Computer

NA

Yes – 103

61

10

81

0.86

0.43

0.57

Results from Applying SVM Train Model on Test

19.

Computer

NA

No – 654

162

8

92

0.95

0.64

0.76

20.

Computer

NA

Yes – 576

158

9

96

0.95

0.62

0.75