Research Article
Link Prediction via Convex Nonnegative Matrix Factorization on Multiscale Blocks
input: //observed network | , //desired rank, scale parameter | output: //predicting matrix | | find community structure | for | if size | divide into intsize equal parts | append each part to | delete | end if | end for | for | if size | for , | if size | | delete | end if | end for | end if | end for | partition matrix into by | for block | do CNMF with rank min | //sum the corresponding entities | end for |
|