Information Technology Reference
In-Depth Information
Table1 illustrates the experimental results of three candidate algorithms. It is clear
to find out that SCPSO outperforms the other three algorithms in most cases. In Karate
network, SCPSO gets a little worse modularity Q than MOGA-NET. The reason is that
clustering process in the framework has combine some nodes, SCPSO may not
retrieval the global best modularity compared to random optimize in the original
network. If splitting some core area increases the modularity index, SCPSO gets lower
modularity index compared to optimize in origin network. In Dolphins, Krebs and
Football, SCPSO outperforms its competitors in both modularity and NMI.
Table 1.
Comparison of Modularity in Real Networks
MODULARITY COMPARISON
NMI COMPARISION
SCPSO
MOGA-NET
GN
SCPSO
MOGA-NET
GN
0.400
0.415
0.380
0.803
0.602
0.692
Karate
Dolphins
0.528
0.505
0.495
0.581
0.506
0.573
Krebs
0.521
0.518
0.502
0.549
0.536
0.530
Football
0.617
0.515
0.577
0.801
0.775
0.762
6
Conclusion
Our goal is to reduce the scale of network and accelerate the convergence in
optimization process. In addition, optimization in the new constructed network has
shown its advantage, and the rationality has been briefly proof. In the optimization
process, a mutation strategy had been proposed to accelerate the convergence. In
comparison with GN, MOGA-Net in synthetic and four real networks, SCPSO exhibits
its advantage. Therefore, the proposed algorithm SCPSO is an effective optimization
algorithm in community detection. Expanding the algorithm to dynamic networks is
our next job.
Acknowledgment.
The authors would like to thank the support of the Technology
Innovation Platform Project of Fujian Province under Grant No. 2009J10027, the Key
Project of Fujian Education Committee under Grant No. JK2012003, the Program of
National Natural Science Foundation of China under Grant No. 60171 009.
References
1.
Fortunato, S.: Community detection in graphs. Physics Reports 486, 75-174 (2010)
2.
Radicchi, F., Castellano, C., Cecconi, F., Loreto, V., Parisi, D.: Defining and identifying
communities in networks. Proceedings of the National Academy of Sciences of the United
States of America 101, 2658-2663 (2004)
3.
Flake, G.W., Lawrence, S., Giles, C.L., Coetzee, F.M.: Self-organization and identification
of web communities. Computer 35, 66-70 (2002)
4.
Ravasz, E., Somera, A.L., Mongru, D.A., Oltvai, Z.N., Barabási, A.-L.: Hierarchical or-
ganization of modularity in metabolic networks. Science 297, 1551-1555 (2002)
