Log-likelihood matrix for the BT–SBM (clustered) model

Builds an T_iter x D matrix of log-likelihood values using cluster labels \(x_i\) and cluster rates \(\lambda_k\). Assumes x_samples and lambda_samples are relabelled consistently (e.g. via inference_helper).

Builds an \(S \times D\) matrix using cluster labels \(x_i\) and cluster rates \(\lambda_k\). Assumes inputs are relabelled consistently or that cluster ids in x_samples[s, ] are in 1..K where K = ncol(lambda_samples).

Usage

make_bt_cluster_loo(w_ij, lambda_samples, x_samples)

make_bt_cluster_loo(w_ij, lambda_samples, x_samples)

Arguments

w_ij

integer/numeric \(n \times n\) wins (i over j).

lambda_samples

numeric \(S \times K\) matrix of cluster rates \(\lambda_k\).

x_samples

integer \(S \times n\) matrix of cluster labels for each item.

Value

A list with:

• ll — T_iter x D matrix of log-likelihoods.

• obs_idx — D x 2 matrix of (i,j) indices defining each column.

A list with:

• ll — \(S \times D\) matrix of log-likelihoods.

• obs_idx — \(D \times 2\) matrix of (i,j) indices defining each column.

Examples

if (FALSE) { # \dontrun{
# After running your clustered sampler and relabeling:
# ll_obj <- make_bt_cluster_loo(w, n, out$lambda_samples_relabel, out$x_samples_relabel)
} # }