[R] a difficulty in boot package

[Uwe Ligges]

Seunghee Baek wrote:
> Hi,
> I have a problem in programming for bootstrapping.
> I don't know why it show the error message.
> Please see my code below:
> #'st' is my original dataset. 
> #functions of 'fml.mlogl','pcopula.fam4','ltd','invltd' are already defined
> 
> boot.OR<-function(data,i)
> {
> E=data[i,]
> ml1<-glm(c_VAsex90_bf ~ trt,family=binomial,data=E)
> ml2<-glm(c_VAsex90_bm ~ trt,family=binomial,data=E)
> marg.covariates<-cbind(rep(1,length(E$trt)),E$trt)
> dep.covariates<-cbind(rep(1,length(E$age_avr)),E$age_avr)
> start<-c(ml1$coef,ml2$coef,0,0)
> fml1<-optim(start,fml.mlogl,control=c(maxit=10000),hessian=F)
> x<-(1+exp(fml1$par[1]))^(-1)
> y<-(1+exp(fml1$par[3]))^(-1)
> b<-exp(fml1$par[5]+fml1$par[6]*43)+1
> p00<-ltd(-log(pcopula.fam4(exp(-invltd(x,b)),exp(-invltd(y,b)),b)),b)
> p1<-exp(fml1$par[1])/(1+exp(fml1$par[1]))
> p2<-exp(fml1$par[3])/(1+exp(fml1$par[3]))
> OR<-p00*(p00+p1+p2-1)/(1-p2-p00)/(1-p1-p00)
> OR
> }
> 
> set.seed(101)
> 
> boot(st,boot.OR,R=500)
> 
> #------------------------------------------------------------#
> I gives following error message:
> Error in fn(par, ...) : object "dep.covariates" not found


Since you have not specified all function I can only guess that you need 
  dep.covariance in function fml.mlogl() in your optim() call. In that 
case you need to pass it as an argument, otherwise it is not in the 
search path given R's scoping rules.

Hence, for example, define

   fml.mlogl <- function(......., dep.covariates)

and call ist with

   optim(........, dep.covariates=dep.covariates)



Best,
Uwe Ligges



> I hope you can help me in this problem.
> 
> Thanks,
> Becky
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.