SSVS.Tuning: MH Random Walk Proposal Tuning for SSVS under Logistic...

In lcw68/G3proj: BIOS 735 Project

Description

Tunes the normal proposal distribution standard deviation for Metropolis Hastings MCMC for the SSVS variable selection method in a Logistic Regression model.

Usage

SSVS.Tuning(Y0, X0, c0, tau0, nMC = 1000, b = 50, seed = 1)

Arguments

Y0	vector of responses
X0	covariate matrix without intercept
c0	parameter for spike and slab prior of beta
tau0	parameter for spike and slab prior of beta
nMC	number of MCMC samples
b	batch size
seed	set seed for random number generation

Details

Runs nMc/b number of batches of size b. In each batch the algorithm calculates the Metropolis acceptance rate and adjusts the proposal standard deviation accordingly. As such, nMC must be a multiple of b.

Value

A nested list of gamma samples, beta samples, beta acceptance rates for each batch, and the proposal standard deviations at the final batch.

Examples

## simulate data;

set.seed(1);
N  = 100;
p  = 10;

X  = matrix(data = rnorm(N*p), nrow=N, ncol=p)
beta_true = c(rep(1,p/2),rep(0,p/2))
eta = X %*% beta_true
pi = exp(eta) / (1 + exp(eta))
Y  = rbinom(N,1,pi)

## fit model;
test1 <- G3proj::SSVS.Tuning(Y0 = Y, X0 = X, c0 = 10,
                             tau0 = 0.4, nMC = 1000, b = 50)

lcw68/G3proj documentation built on Dec. 21, 2021, 9:46 a.m.