Skip to main content

Table 3 Execution time (in seconds) of the ADVNDS, Leiden, Louvain, Combo, and GNNS algorithms over the classic network examples

From: Graph neural network inspired algorithm for unsupervised network community detection

No

Name

ADVNDS

Leiden

Louvain

Combo

GNN100

GNN2500

GNN25000

1

karate

2.5

0.02

0.01

0.02

0.04

0.23

4.81

2

chesapeake

3.3

0.02

0.01

0.03

0.03

0.23

4.86

3

dolphins

0.0

0.07

0.01

0.06

0.03

0.26

5.05

4

lesmis

0.0

0.03

0.01

0.20

0.03

0.26

5.28

5

polbooks

0.0

0.08

0.01

0.42

0.03

0.27

5.26

6

adjnoun

2.5

0.12

0.03

0.65

0.03

0.30

5.61

7

football

10.8

0.09

0.02

0.41

0.03

0.35

6.20

8

jazz

30.5

0.23

0.16

0.88

0.07

0.67

6.10

9

C.Elegans N

1.5

0.46

0.29

4.64

0.10

0.71

7.50

10

US Airports1997

–

0.58

0.19

6.08

0.08

0.88

6.97

11

C.Elegans M

16.6

1.00

0.26

24.73

0.08

1.38

12.93

12

email

297.3

2.72

0.72

141.41

0.28

2.71

31.24

13

polblogs

1.2

1.62

0.72

56.63

0.75

15.58

195.01

14

US Airports2010

–

2.06

1.01

170.47

0.85

19.08

237.23

15

power

16439.4

6.28

1.12

2200.80

1.43

39.60

491.33

16

PGPgiantcompo

18000.0

11.41

2.59

25773.10

3.32

101.95

1256.97

17

Krong500slogn16

17610.5

5983.00

298.22

–

45.78

1061.30

13736.06

  1. The best results are in bold