BALM
BALM gives posterior uncertainty for topology, weights, sparsity coupling, templates, and subject-specific mixture weights. BALM combines continuous mixed membership, low-rank latent score templates, and a hurdle likelihood for zero-inflat…
1 sources - 4 claims
BALM gives posterior uncertainty for topology, weights, sparsity coupling, templates, and subject-specific mixture weights. BALM combines continuous mixed membership, low-rank latent score templates, and a hurdle likelihood for zero-inflated connectomes. BALM is designed to learn population-level connectivity motifs and quantify each subject's degree of expression of those motifs. BALM is applicable beyond connectomics to other sparse weighted replicated networks.