In recent years we have witnessed that the crime presents the gang characteristic. In many cases, we are able to obtain the communication messages of the suspicious conspirators. However, it is rather difficult to find out most conspirators with limited information, especially when the number of suspicious conspirators is large. In fact, to find out the conspirators and non-conspirators is equivalent to separate the network into two communities. There are some successful algorithms to separate an unweighted network into several communities. But they are of little use to our problem since the network of our problem is a weighted graph. In order to better solve the problem, we creatively convert the network into a circuit and use Kirchhoff's Voltage Law (KVL) to figure out the result by iteration. This method meets with the features of our problem very well. And it turns out to be reasonable and effective theoretically considering its strict process of mathematical derivation.
[1]
Deng Cai,et al.
Topic modeling with network regularization
,
2008,
WWW.
[2]
Fang Wu,et al.
Finding communities in linear time: a physics approach
,
2003,
ArXiv.
[3]
Michael I. Jordan,et al.
Latent Dirichlet Allocation
,
2001,
J. Mach. Learn. Res..
[4]
M E J Newman,et al.
Finding and evaluating community structure in networks.
,
2003,
Physical review. E, Statistical, nonlinear, and soft matter physics.
[5]
Brian W. Kernighan,et al.
An efficient heuristic procedure for partitioning graphs
,
1970,
Bell Syst. Tech. J..