Estimation of the Hybrid DSBM model via the CEM algorithm

Estimation of the Hybrid DSBM model via the CEM algorithm

Usage

est_pred_sbm_dyn_decoding(
  Y,
  k,
  start = 0,
  tol = 10^-10,
  maxit = 100,
  Tau = NULL
)

Arguments

Y

3D binary array (n x n x TT) representing the temporal adjacency matrix.

k

Number of latent blocks/classes.

start

Initialization strategy (0 = K-means, 1 = random, 2 = custom Tau).

tol

Convergence tolerance for the log-likelihood.

maxit

Maximum number of CEM iterations.

Tau

Initial membership array (required if start = 2).

Value

List containing parameter estimates, likelihood, and classifications (cl and clv).