Movatterモバイル変換


[0]ホーム

URL:


frailtySurv

frailtySurv is an R package for simulating and fitting semiparametricshared frailty models.

Installation

For the latest stable version, install from CRAN:

install.packages("frailtySurv")

The development version can be installed from github usingdevtools.

devtools::install_github("vmonaco/frailtySurv")

Example

The following code shows how to generate data and fit the model.

set.seed(1234)library(frailtySurv)#> Loading required package: survival#> Welcome to frailtySurv v1.3.3dat<-genfrail(N=200,K=2,beta=c(log(2),log(3)),frailty="gamma",theta=2,censor.rate=0.35,Lambda_0=function(t,tau=4.6,C=0.01) (C*t)^tau)fit<-fitfrail(Surv(time, status)~ Z1+ Z2+cluster(family),dat=dat,frailty="gamma")
fit#> Call: fitfrail(formula = Surv(time, status) ~ Z1 + Z2 + cluster(family),#>     dat = dat, frailty = "gamma")#>#>      Covariate     Coefficient#>             Z1           0.654#>             Z2           1.006#>#> Frailty distribution   gamma(1.802), VAR of frailty variates = 1.802#> Log-likelihood         -1575.752#> Converged (method)     10 iterations, 3.24 secs (maximized log-likelihood)

Parameter traces are given by

plot(fit,type="trace")

The estimated cumulative baseline hazard is given by

plot(fit,type="cumhaz")

The results can be compared to other estimation techniques.

library(survival)library(frailtypack)
coxph(Surv(time, status)~ Z1+ Z2+frailty.gamma(family),data=dat)#> Call:#> coxph(formula = Surv(time, status) ~ Z1 + Z2 + frailty.gamma(family),#>     data = dat)#>#>                           coef se(coef)      se2    Chisq  DF       p#> Z1                      0.6651   0.0941   0.0852  49.9190   1 1.6e-12#> Z2                      1.0282   0.1046   0.0941  96.5685   1 < 2e-16#> frailty.gamma(family)                            475.6283 150 < 2e-16#>#> Iterations: 7 outer, 58 Newton-Raphson#>      Variance of random effect= 1.88   I-likelihood = -1319.6#> Degrees of freedom for terms=   0.8   0.8 149.7#> Likelihood ratio test=576  on 151 df, p=<2e-16  n= 400frailtyPenal(Surv(time, status)~ Z1+ Z2+cluster(family),data=dat,n.knots=10,kappa=2)#>#> Be patient. The program is computing ...#> The program took 0.22 seconds#> Call:#> frailtyPenal(formula = Surv(time, status) ~ Z1 + Z2 + cluster(family),#>     data = dat, n.knots = 10, kappa = 2)#>#>#>   Shared Gamma Frailty model parameter estimates#>   using a Penalized Likelihood on the hazard function#>#>       coef exp(coef) SE coef (H) SE coef (HIH)       z          p#> Z1 0.67236   1.95885   0.0994732     0.0994732 6.75921 1.3875e-11#> Z2 1.04884   2.85433   0.1191210     0.1191210 8.80481 0.0000e+00#>#>     Frailty parameter, Theta: 1.92378 (SE (H): 0.319435 ) p = 8.5895e-10#>#>       penalized marginal log-likelihood = -1378.15#>       Convergence criteria:#>       parameters = 0.000362 likelihood = 0.000201 gradient = 2.44e-08#>#>       LCV = the approximate likelihood cross-validation criterion#>             in the semi parametrical case     = 3.48287#>#>       n= 400#>       n events= 256  n groups= 200#>       number of iterations:  13#>#>       Exact number of knots used:  10#>       Value of the smoothing parameter:  2, DoF:  11.96

Clone and build

You can clone the repository and build the project from source usingRStudio. To create a project in RStudio from this repository, you musthave both RStudio and git installed.

See theR packages book

Build everything

To load the package (simulate installing and loading withlibrary(frailty)):

devtools::load_all()

or shortuct:Ctrl+Shift+L

Use this command to build a source package (binary = FALSE bydefault)

devtools::build()

Other usefull commands and shortcuts:

Build vignettes

Build the vignettes with the command:

devtools::build_vignettes()

Compiled vignettes will reside ininst/doc. Inparticular, see frailtySurv.pdf

git pull

Use RStudio to pull the latest code if you’ve already cloned therepository. In the “git” pane, look for the Pull button. You can alsoview the commit history from there and see the diff between modifiedfiles, if any.


[8]ページ先頭

©2009-2025 Movatter.jp