MYULA: MYULA transition step

In qhengncsu/BNRPxMCMC: Bayesian l1 Trend Filtering via Proximal MCMC

Description

MYULA transition step

Usage

MYULA(
  theta,
  sigma2,
  wimax,
  lambda,
  gradf,
  prox,
  delta = 1,
  t = 1,
  MH_correct = F,
  logpi = NULL
)

Arguments

theta	current theta iterate (vector)
sigma2	current sigma2 iterate
wimax	the maximum number of observations at a given location
lambda	lambda parameter of Moreau envelope
gradf	function handle that computes the gradient of f
prox	function handle of the proximal operator of g
delta	percentage of max step size to use
t	number of theta sampling steps to carry out sigma2
MH_correct	whether to use MH correction or not
logpi	the target conditional density of theta

Value

the new theta iterate (vector); if using MH correction, return (thetanew,accept)

qhengncsu/BNRPxMCMC documentation built on Dec. 31, 2020, 2:10 a.m.