dfpSpiLasSte: Performs one iteration of DFP for the Spike & Lasso prior.

In Rene-Gutierrez/DynParRegReg:

Description

This function performs one step of the DFP with the Bayesian Lasso prior. The inputs are described in Algorithm 1 and the Supplementary Materials of Guhaniyogi & Gutierrez.

Usage

dfpSpiLasSte(
  XX,
  Xy,
  yy,
  sN,
  c,
  hb,
  ht,
  hl,
  hs,
  hd,
  hg,
  hlsh = 1,
  hlsc = 1,
  hds1 = 1,
  hds2 = 1,
  nmcmc
)

Arguments

XX	Sufficient statistic X'X.
Xy	Sufficient statistic X'y.
yy	Sufficient statistic y'y.
sN	Number of Observations.
c	Spike variance parameter.
hb	Point estimate of the coefficients.
ht	Point estimate for the local shrinking parameters.
hl	Point estimate for the global shrinking parameter.
hs	Point estimates of the variance of the error.
hd	Point estimate for the sparsity parameter.
hg	Point estimate for the indicator functions.
hlsh	Hyper-prior for the lambda parameter shape.
hlsc	Hyper-prior for the lambda parameter rate.
hds1	Hyper-prior for the sparsity parameter.
hds2	Hyper-prior for the sparsity parameter.
nmcmc	Number of samples of the parameters.

Value

A list containing samples for the model parameters.

sb

A matrix with samples for coefficients, 1 sample per row.

st

A matrix with samples for local shrinkage parameters, 1 sample per row.

ss

Samples for sigma.

sl

Samples for the global shrinkage parameter.

sd

Samples for the sparsity parameter.

sg

Samples for the Lasso indicators.

Author(s)

Rene Gutierrez Marquez

Rene-Gutierrez/DynParRegReg documentation built on Dec. 18, 2021, 9:57 a.m.