compute_ABC_cpp: Perform ABC sampling using the stick breaking procedure,...

In lgaborini/rdirdirgamma: Generate samples from Dirichlet-Dirichlet-Gamma model, and perform ABC.

Description

Perform ABC sampling using the stick breaking procedure, returning the acceptance ratio. Similar to sample_ABC_rdirdirgamma_beta_cpp() but also performs the acceptance step.

Usage

compute_ABC_cpp(
  n_sample,
  m_sample,
  alpha_0,
  beta_0,
  nu_0,
  mtx_obs,
  summarize_eps,
  reps,
  p_norm,
  use_optimized_summary,
  return_distances = FALSE
)

Arguments

n_sample	hyperparameters that are used to generate data: number of samples per source
m_sample	hyperparameters that are used to generate data: number of sources
alpha_0	hyperparameters that are used to generate data
beta_0	hyperparameters that are used to generate data
nu_0	hyperparameters that are used to generate data
mtx_obs	the observed data matrix
summarize_eps	ABC thresholds: as many as summary statistics
reps	number of ABC samples to generate
p_norm	exponent of the L^p norm (can be Inf) (default: 2)
use_optimized_summary	if TRUE, return the optimized summary statistics (mean, sd, kurtosis, skewness), else standard (mean, sd)
return_distances	if TRUE, also return distances for all samples

Value

a list with components:

• n_accepted: number of accepted samples

• accept_ratio: the acceptance ratio, where 1 means that all max_iter samples were accepted.

• d_ABC: a (max_iter x n_summary) matrix of distances (if return_distances is TRUE)

See Also

Other ABC functions: compute_distances_gen_obs_cpp(), generate_acceptable_data_cpp(), sample_ABC_rdirdirgamma_beta_cpp()

lgaborini/rdirdirgamma documentation built on March 6, 2021, 3:05 p.m.