Table 3 Precision-ranking of the different network.

From: A perturbation-based framework for link prediction via non-negative matrix factorization

 

C. elegant

Email

Karate

Word

Jazz

USAir

Yeast

PB

NS

Router

Power

Baydry

School

SmaGri

SW

mean

NMF-A2*(2)

1

3

4

4

4

2

1

1

7

1

6

2

2

1

1

2.67

SPM*(2)

5

5

3

1

1

3

5

2

4

2

2

1

1

3

3

2.73

NMF-D2*(2)

2

4

5

5

5

4

2

3

8

6

11

5

9

2

2

4.87

CRA

8

1

7

15

6

6

7

4

3

7

8

11

5

4

8

6.67

NMF-A1*(2)

3

7

2

14

3

7

3

10

6

14

12

3

3

11

6

6.93

NMF-D1*(2)

4

6

1

13

2

11

4

9

11

16

13

4

4

10

5

7.53

AA

9

2

13

8

10

5

11

8

2

11

9

15

16

5

10

8.93

RA

10

9

11

10

8

1

12

12

1

15

10

14

8

6

9

9.07

Katz(1)

11

11

10

7

17

10

8

6

10

8

1

10

17

7

7

9.33

LP(1)

12

10

9

6

15

9

9

5

9

9

4

12

19

8

11

9.80

NMF1*

6

12

6

12

7

14

6

13

12

12

14

6

7

12

12

10.07

CN

13

8

12

9

14

8

10

7

5

10

5

13

15

9

13

10.07

NMF2*

7

13

8

11

13

12

13

11

17

4

15

7

14

14

4

10.87

ACT

15

18

15

3

19

13

23

14

18

3

7

9

18

15

14

13.60

TSCN(1)

20

20

14

20

23

16

14

16

20

5

3

17

6

16

16

15.07

PA

14

19

16

2

20

15

16

15

22

13

23

8

23

13

15

15.60

Jaccard

17

14

21

18

11

18

18

18

15

18

20

20

10

19

18

17.00

Salton

19

17

20

21

9

20

17

20

14

17

16

19

12

18

17

17.07

HDI

16

15

18

17

16

17

20

17

13

21

21

22

13

21

20

17.80

Sorenson

18

16

22

19

12

19

19

19

16

19

19

21

11

20

19

17.93

HPI

21

21

17

16

18

21

15

21

19

20

22

16

20

17

23

19.13

LHN

22

22

19

23

21

22

21

22

21

22

17

18

21

22

21

20.93

LHNII(2)

23

23

23

22

22

23

22

23

23

23

18

23

22

23

22

22.33

  1. The last column is the mean precision-ranking across all the networks. Different methods are presented in increasing order of mean precision-ranking. The local methods are in standard character while the global methods are in italic. The name of the best local method and the name of the best global method in the first column are boldface. Corresponding, the mean precision-ranking value of the best local method and the mean precision-ranking value of the best global method in the last column also are boldface. Number in bracket closed to an method denotes the number of tuning parameters. The best result achieved by global methods and the best result achieved by local methods on each network are boldface. Methods with an asterisk like * denote methods based on inference and methods without an asterisk denote methods based on a paradigm (in the sense that are model-based).
Back to article page