hierarchical_fusion_RIS_TA_BLR: Time-adapting Hierarchical Rejection-Importance Sampling...

In rchan26/BayesLogitFusion: Monte Carlo Fusion for Bayesian Logistic Regression

Description

Time-adapting Hierarchical Rejection-Importance Sampling Monte Carlo Fusion for Bayesian Logistic Regression model

Usage

hierarchical_fusion_RIS_TA_BLR(
  N_schedule,
  dim,
  y_split,
  X_split,
  prior_means,
  prior_variances,
  global_T,
  m_schedule,
  C,
  power,
  precondition = FALSE,
  L,
  base_samples,
  ESS_threshold = 0.5,
  seed = NULL,
  n_cores = parallel::detectCores()
)

Arguments

N_schedule	vector of length (L-1), where N_schedule[l] is the number of samples per node at level l
dim	dimension of the predictors (= p+1)
y_split	list of length C, where y_split[[c]] is the y responses for sub-posterior c
X_split	list of length C, where X_split[[c]] is the design matrix for sub-posterior c
prior_means	prior for means of predictors
prior_variances	prior for variances of predictors
global_T	time T for time-adapting fusion algorithm
m_schedule	vector of length (L-1), where m_schedule[l] is the number of samples to fuse for level l
C	number of sub-posteriors at the base level
power	exponent of target distribution
precondition	boolean value determining whether or not a preconditioning matrix is to be used
L	total number of levels in the hierarchy
base_samples	list of length C, where samples_to_fuse[c] containg the samples for the c-th node in the level
ESS_threshold	number between 0 and 1 defining the proportion of the number of samples that ESS needs to be lower than for resampling (i.e. resampling is carried out only when ESS < N*ESS_threshold)
seed	seed number - default is NULL, meaning there is no seed
n_cores	number of cores to use

Value

A list with components:

samples

list of length (L-1), where samples[[l]][[i]] are the samples for level l, node i

weighted_samples

list of length (L-1), where weighted_samples[[l]][[i]] are the weighted samples for level l, node i

normalised_weights

list of length (L-1), where normalised_weights[[l]][[i]] is the normalised weights for level l, node i

ESS

list of length (L-1), where ESS[[l]][[i]] is the ESS for samples in level l, node i

x_samples

list of length (L-1), where x_samples[[l]][[i]] are the samples from first fusion step (rho step) for level l, node i

rho_acc

list of length (L-1), where rho_acc[[l]][i] is the acceptance rate for first fusion step for level l, node i

time

list of length (L-1), where time[[l]] is the run time for level l, node i

resampled

list of length (L-1), where resampled[[l]][[i]] is a boolean value to indicate if samples in level, node i were resampled

y_inputs

input y data for each level and node

X_inputs

input X data for each level and node

C_inputs

vector of length (L-'), where C_inputs[l] is the number of sub-posteriors at level l+1 (the input for C to get to level l)

sub_posterior_weight_inputs

list of length (L), where sub_posterior_weight_inputs[[l]] is the input for the sub-posterior weights for level l

diffusion_times

vector of length (L-1), where diffusion_times[l] are the times for fusion in level l

power

exponent of target distributions in the hierarchy

precondition_matrices

pre-conditioning matricies that were used

rchan26/BayesLogitFusion documentation built on June 13, 2020, 5:03 a.m.