ea_BRR_DL_PT: Diffusion probability for the Exact Algorithm for Langevin...

In rchan26/hierarchicalFusion: Divide-and-Conquer Monte Carlo Fusion

Diffusion probability for the Exact Algorithm for Langevin diffusion for Bayesian robust regression

Description

Simulate Langevin diffusion using the Exact Algorithm where target is the posterior for a robust regression model with Gaussian priors

Usage

ea_BRR_DL_PT(
  dim,
  x0,
  y,
  s,
  t,
  data,
  transformed_design_mat,
  nu,
  sigma,
  prior_means,
  prior_variances,
  C,
  precondition_mat,
  transform_mats,
  diffusion_estimator,
  beta_NB = 10,
  gamma_NB_n_points = 2,
  logarithm
)

Arguments

dim	dimension of the predictors (= p+1)
x0	start value (vector of length dim)
y	end value (vector of length dim)
s	start time
t	end time
data	list of length 4 where data[[c]]$y is the vector for y responses and data[[c]]$X is the design matrix for the covariates for sub-posterior c
nu	degrees of freedom in t-distribution
sigma	scale parameter in t-distribution
prior_means	prior for means of predictors
prior_variances	prior for variances of predictors
C	overall number of sub-posteriors
precondition_mat	precondition matrix
transform_mats	list of transformation matrices where transform_mats$to_Z is the transformation matrix to Z space and transform_mats$to_X is the transformation matrix to X space
diffusion_estimator	choice of unbiased estimator for the Exact Algorithm between "Poisson" (default) for Poisson estimator and "NB" for Negative Binomial estimator
beta_NB	beta parameter for Negative Binomial estimator (default 10)
gamma_NB_n_points	number of points used in the trapezoidal estimation of the integral found in the mean of the negative binomial estimator (default is 2)
logarithm	logical value to determine if log probability is returned (TRUE) or not (FALSE)

Value

acceptance probability of simulating Langevin diffusion where target is the posterior for a robust regression model with Gaussian priors

rchan26/hierarchicalFusion documentation built on Sept. 11, 2022, 10:30 p.m.