psbcGroup (version 1.7)

aftGL_LT: Function to Fit the Penalized Parametric Bayesian Accelerated Failure Time Model with Group Lasso Prior for Left-Truncated and Interval-Censored Data

Description

Penalized parametric Bayesian accelerated failure time model with group lasso prior is implemented to analyze left-truncated and interval-censored survival data with high-dimensional covariates.

Usage

aftGL_LT(Y, X, XC, grpInx, hyperParams, startValues, mcmcParams)

Value

aftGL_LT returns an object of class aftGL_LT.

Arguments

Y

Outcome matrix with three column vectors corresponding to lower and upper bounds of interval-censored data and left-truncation time

X

Covariate matrix \(p\) covariate vectors from n subjects. It is of dimension \(n\times p\).

XC

Matrix for confound variables: \(q\) variable vectors from n subjects. It is of dimension \(n\times q\).

grpInx

a vector of \(p\) group indicator for each variable

hyperParams

a list containing hyperparameter values in hierarchical models: (a.sigSq, a.sigSq): hyperparameters for the prior of \(\sigma^2\); (mu0, h0): hyperparameters for the prior of \(\mu\); (v): hyperparameter for the prior of \(\beta_C\).

startValues

a list containing starting values for model parameters. See Examples below.

mcmcParams

a list containing variables required for MCMC sampling. Components include, numReps, total number of scans; thin, extent of thinning; burninPerc, the proportion of burn-in. See Examples below.

Author

Kyu Ha Lee, Harrison Reeder

References

Reeder, H., Haneuse, S., Lee, K. H. (2024+). Group Lasso Priors for Bayesian Accelerated Failure Time Models with Left-Truncated and Interval-Censored Data. under review

See Also

VS

Examples

if (FALSE) {

data(survData)
X <- survData[,c(4:5)]
XC <- NULL

n <- dim(survData)[1]
p <- dim(X)[2]
q <- 0

c0 <- rep(0, n)
yL <- yU <- survData[,1]
yU[which(survData[,2] == 0)] <- Inf
Y <- cbind(yL, yU, c0)

grpInx <- 1:p
K <- length(unique(grpInx))

#####################
## Hyperparameters

a.sigSq= 0.7
b.sigSq= 0.7

mu0 <- 0
h0 <- 10^6

v = 10^6

hyperParams <- list(a.sigSq=a.sigSq, b.sigSq=b.sigSq, mu0=mu0, h0=h0, v=v)

###################
## MCMC SETTINGS

## Setting for the overall run
##
numReps    <- 100
thin    <- 1
burninPerc <- 0.5

## Tuning parameters for specific updates
##

L.beC <- 50
M.beC <- 1
eps.beC <- 0.001

L.be <- 100
M.be <- 1
eps.be <- 0.001

mu.prop.var    <- 0.5
sigSq.prop.var    <- 0.01

##

mcmcParams <- list(run=list(numReps=numReps, thin=thin, burninPerc=burninPerc),
tuning=list(mu.prop.var=mu.prop.var, sigSq.prop.var=sigSq.prop.var,
L.beC=L.beC, M.beC=M.beC, eps.beC=eps.beC,
L.be=L.be, M.be=M.be, eps.be=eps.be))

#####################
## Starting Values

w        <- log(Y[,1])
mu     <- 0.1
beta     <- rep(2, p)
sigSq    <- 0.5
tauSq <- rep(0.4, p)
lambdaSq <- 100
betaC     <- rep(0.11, q)

startValues <- list(w=w, beta=beta, tauSq=tauSq, mu=mu, sigSq=sigSq,
lambdaSq=lambdaSq, betaC=betaC)

fit <- aftGL_LT(Y, X, XC, grpInx, hyperParams, startValues, mcmcParams)

}