SBM Ising

This repository uses Ising model (simulated annealing) to recover labels of stochastic block model.

Note

The code in this repository can only recover symmetric SBM. Symmetric SBM means that the communities have equal size.

The SBM generator is $\textrm{SBM}(n, k, \frac{a \log n}{n}, \frac{b \log n}{n})$.

Sample code

from sbmising import SIBM, sbm_graph
G = sbm_graph(100, 2, 16, 4)
X = SIBM(G, k=2)
print(X)

Reference

[1] Zhao, Feng, Min Ye, and Shao-Lun Huang. "Exact Recovery of Stochastic Block Model by Ising Model." Entropy 23.1 (2021): 65.