Effective Semisupervised Community Detection Using Negative Information

Algorithm 1

Semisupervised community detection using negative information.

   input: adjacency matrix , the initial label matrix , , the constants and

   output: The TL of all the nodes

()   construct the weight matrix by (1).

()   for

()    If has the TL

()    construct and by (2) and (3), respectively.

()    If has the NLs

()    construct and by (4) and (5), respectively.

()    If is unlabeled node

()    construct and by (6) and (7), respectively.

()   iterate (9) until convergence.

() output the labels of each node by .