Generalization of The Mean-field Method for Power-law Distributions

摘要

In this paper, the generalized mean-field method has applied to power-law distributions. We have shown that this method greatly reduces the number of operations to calculate expected values of the distributions when the projection is appropriately selected. We carried out this method to binary-state power-law distributions and obtained less errors of expected values compared with those obtained by the naive mean-field method. It is an important future work to apply this method to real-world systems.