predictive {bayesSurv}

Compute predictive quantities based on a Bayesian survival regression model fitted using bayessurvreg1 function.

Description

This function runs additional McMC to compute predictive survivor and hazard curves and predictive event times for specified values of covariates.

Firstly, the function bayessurvreg1 has to be used to obtain a sample from the posterior distribution of unknown quantities.

Directly, posterior predictive quantiles and means of asked quantities are computed and stored in files.

Function predictive.control serves only to perform some input checks inside the main function predictive.

Usage

predictive(formula, random, time0 = 0, data = parent.frame(),
     grid, type = "mixture", subset, na.action = na.fail,
     quantile = c(0, 0.025, 0.5, 0.975, 1),                       
     skip = 0, by = 1, last.iter, nwrite, only.aver = FALSE,
     predict = list(Et=TRUE, t=FALSE, Surv=TRUE, hazard=FALSE, cum.hazard=FALSE),
     store = list(Et=TRUE, t = FALSE, Surv = FALSE, hazard = FALSE, cum.hazard=FALSE),
     Eb0.depend.mix = FALSE,
     dir, toler.chol = 1e-10, toler.qr = 1e-10)

predictive.control(predict, store, only.aver, quantile)

Arguments

formula

the same formula as that one used to sample from the posterior distribution of unknown quantities by the function bayessurvreg1.

random

the same random statement as that one used to sample from the posterior distribution of unknown quantities by the function bayessurvreg1.

time0

starting time for the survival model. This option is used to get correct hazard function in the case that the original model was \log(T-time0) = \dots.

data

optional data frame in which to interpret the variables occuring in the formulas. Usually, you create a new data.frame similar to that one used to obtain a sample from the posterior distribution. In this new data.frame, put covariate values equal to these for which predictive quantities are to be obtained. If cluster statement was used, assign a unique cluster identification to each observation. Response variable and a censoring indicator may be set to arbitrary values. They are only used in formula but are ignored for computation.

grid

a list of length as number of observations in data or a vector giving grids of values where predictive survivor functions, hazards, cumulative hazards are to be evaluated. If it is a vector, same grid is used for all observations from data. Not needed if only predict$t or predict$Et are TRUE. If time0 is different from zero your grid should start at time0 and not at zero.

type

a character string giving the type of assumed error distribution. Currently, valid are substrings of "mixture". In the future, "spline", "polya.tree" might be also implemented.

subset

subset of the observations from the data to be used. This option will normally not be needed.

na.action

function to be used to handle any NAs in the data. The user is discouraged to change a default value na.fail.

quantile

a vector of quantiles that are to be computed for each predictive quantity.

skip

number of rows that should be skipped at the beginning of each *.sim file with the stored sample.

by

additional thinning of the sample.

last.iter

index of the last row from *.sim files that should be used. If not specified than it is set to the maximum available determined according to the file mixmoment.sim.

nwrite

frequency with which is the user informed about the progress of computation (every nwriteth iteration count of iterations change).

only.aver

if TRUE only posterior predictive mean is computed for all quantities and no quantiles.

predict

a list of logical values indicating which predictive quantities are to be sampled. Components of the list:

Et

predictive expectations of survivor times

t

predictive survivor times

Surv

predictive survivor functions

hazard

predictive hazard functions

cum.hazard

predictive cumulative hazard functions

store

a list of logical values indicating which predictive quantities are to be stored in files as ‘predET*.sim’, ‘predT*.sim’, ‘predS*.sim’, ‘predhazard*.sim’, ‘predcumhazard*.sim’. If you are interested only in posterior means or quantiles of the predictive quantities you do not have to store sampled values. Posterior means and quantiles are stored in files ‘quantET*.sim’, ‘quantT*.sim’, ‘quantS*.sim’, ‘quanthazard*.sim’, ‘quantpredhazard*.sim’.

Eb0.depend.mix

a logical value indicating whether the mean of the random intercept (if included in the model) was given in a hierarchical model as an overall mean of the mixture in the error term. With FALSE (default) you have the same model as that one described in an accompanying paper. An ordinary user is discouraged from setting this to TRUE.

dir

a string giving a directory where previously simulated values were stored and where newly obtained quantities will be stored. On Unix, do not use ‘~/’ to specify your home directory. A full path must be given, e.g. ‘/home/arnost/’.

toler.chol

tolerance for the Cholesky decomposition.

toler.qr

tolerance for the QR decomposition.

Value

An integer which should be equal to zero if everything ran fine.

Author(s)

Arnošt Komárek [e-mail]

References

Komárek, A. (2006). Accelerated Failure Time Models for Multivariate Interval-Censored Data with Flexible Distributional Assumptions. PhD. Thesis, Katholieke Universiteit Leuven, Faculteit Wetenschappen.

Komárek, A. and Lesaffre, E. (2007). Bayesian accelerated failure time model for correlated interval-censored data with a normal mixture as an error distribution. Statistica Sinica, 17, 549 - 569.

Examples

## See the description of R commands for
## the models described in
## Komarek (2006),
## Komarek and Lesaffre (2007).
##
## R commands available
## in the documentation
## directory of this package as
## - ex-cgd.R and
##   https://www2.karlin.mff.cuni.cz/~komarek/software/bayesSurv/ex-cgd.pdf
##
## - ex-tandmobMixture.R and
##   https://www2.karlin.mff.cuni.cz/~komarek/software/bayesSurv/ex-tandmobMixture.pdf
##