Robustness & reliability

Real network data are noisy and incomplete, and Infomap always returns some partition. This section covers how to keep sparse or incompletely sampled data from fragmenting into small, spurious modules.

• Incomplete data and regularization: when links are missing, standard Infomap overfits into many small spurious modules; the Bayesian regularized map equation adds a prior network so that only well-supported communities survive.