BLRM.Tuning: Bayesian Logistic Regression Model (BLRM) Tuning for Proposal...

In lcw68/G3proj: BIOS 735 Project

Description

Performs hyper parameter tuning of proposal standard deviation that can be used in Bayesian Logistic Regression Model training with user specified parameters and data

Usage

BLRM.Tuning(Y0, X0, PriorVar, nMC = 10000, b = 50, seed = 1)

Arguments

Y0	vector of responses
X0	covariate matrix
PriorVar	variance of prior distribution of beta
nMC	number of MCMC samples, should be the multiple of b
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 beta samples, beta acceptance rates and proposal SD for each 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)

## Compute optimal proposal SD;
test1 <- BLRM.Tuning(Y0 = Y, X0 = X, PriorVar = 1000,
                     nMC = 100, b = 10)

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