sample_lambda_and_zeros: Sampling step for lambda and binary, sparsity-inducing zero...

In EandrewJones/mmBPFA: Multi-Modal Beta-Process Factor Analysis

Description

Sampling step for lambda and binary, sparsity-inducing zero matrix

Usage

sample_lambda_and_zeros(
  x,
  lambda,
  zeros,
  omega,
  alpha,
  gamma_k,
  dc,
  ibp_a,
  ibp_b,
  tau,
  sparse,
  infinite
)

Arguments

x	A multivariate Gaussian copula matrix of dimenion N x P.
lambda	A standard multivariate normal matrix of factor loadings of dimension P x K.
zeros	A binary, sparsity inducing matrix of dimenion P x K.
omega	A standard multivariate normal matrix of factor scores of dimenion N x K.
alpha	A numeric vector of item-intercepts of length P.
gamma_k	A numeric vector of factor precisions of length K.
dc	A integer vector of counts for the number of times each dimensions has been sampled.
ibp_a	A scalar double, Indian Buffet Process beta parameter.
ibp_b	A scalar double, Indian Buffet Process beta parameter.
tau	A tunable hyperparameter for the poisson draws of new dishes to be sampled.
sparse	A boolean indicating whether to include sparsity-inducing prior.
infinite	A boolen indicating whether to sample new potential dishes at each step.

Value

zeros A binary, sparsity inducing matrix of dimenion P x K.

lambda A standard multivariate normal matrix of factor loadings of dimension P x K.

dc A integer vector of counts for the number of times each dimensions has been sampled.

EandrewJones/mmBPFA documentation built on Feb. 14, 2021, 11:17 p.m.