surv.resid: Cox-Snell residuals
In sa4khan/AFTjmr: Recurrent event analysis and joint modeling using accelerated failure time (AFT) models

Description Usage Arguments Details Value Author(s) References See Also Examples

Cox-Snell residuals of the survreg.aft fit.

1 2	surv.resid(fit, plot = FALSE, xlim = NULL, ylim = NULL, main = NULL)

`fit`	`survreg.aft` fit.
`plot`	returns residual plot if `TRUE` (default is `FALSE`).
`xlim`	x limits (x1, x2) of the plot (optional).
`ylim`	y limits of the plot (optional).
`main`	main title of the plot (optional).

Plot of the Cox-Snell (hazard-based) residuals should be roughly a straight line with unit slope when the model is adequate. See Section 6.2 of Lawless and Section 3.7.3 of Cook and Lawless for details.

residuals Cox-Snell residuals.

plot residual plot if plot=TRUE.

Shahedul Khan <khan@math.usask.ca>

Cook RJ and Lawless J, The statistical analysis of recurrent events, Springer, 2007.

Lawless J, Statistical models and methods for lifetime data, Wiley, 2003.

survreg.aft, LR.test.

# Example 1: Recurrent Event Analysis
  library(frailtypack)
  data(readmission)
# Recoding sex and chemo
  sex<-factor(2-as.numeric(readmission$sex),
       labels=c("Female","Male"))
  chemo<-factor(as.numeric(readmission$chemo)-1,
         labels=c("NonTreated","Treated"))
  Data<-readmission
  Data$sex<-sex
  Data$chemo<-chemo
  fit.gg<-survreg.aft(c(t.start,t.stop,event)~sex+chemo+charlson,
      Data=Data,model="ggamma")
  fit.w<-survreg.aft(c(t.start,t.stop,event)~sex+chemo+charlson,
      Data=Data,model="weibull")
  surv.resid(fit.w)
  par(mfrow=c(1,2))
  surv.resid(fit.w,plot=TRUE,xlim=c(0,4.75),ylim=c(0,4.75),main="Weibull")
  surv.resid(fit.gg,plot=TRUE,xlim=c(0,4.75),ylim=c(0,4.75),main="Generalized Gamma")

# Example 2: Non-recurrent Data (Right Censored)
  library(survival)
  fit.ll<-survreg.aft(c(time,status)~karno+diagtime+age+prior+celltype+trt,
      Data=veteran,model="llogistic")
  surv.resid(fit.ll,plot=TRUE,main="Log-logistic")