[R] Calling procedures

Rui Barradas ru|pb@rr@d@@ @end|ng |rom @@po@pt
Mon Jan 25 09:59:43 CET 2021


Hello,

I am not seeing errors, except that you haven't posted the code for 
pbivnorm. Do you have a variable named T somewhere? Don't abbreviate 
TRUE to T in more complex code. Or FALSE to F.

And both functions mycdf and mycdf2 could be simplified.


mycdf <- function(q,logistic=FALSE){
   # *********************************************
   # Univariate CDF: normal or logistic
   # *********************************************
   if(!logistic){
     pnorm(q)
   } else {
     plogis(q)
   }
}

mycdf2 <- function(x,y,rho,logistic=FALSE){
   # *********************************************
   # Calling bivariate CDF: normal or logistic
   # *********************************************
   if(!logistic){
     pbivnorm(x,y,rho,recycle=TRUE)
   } else {
     pbivlogis(x,y,rho)
   }
}

As one-liners:


mycdf <- function(q,logistic=FALSE){
   # *********************************************
   # Univariate CDF: normal or logistic
   # *********************************************
   if(logistic) plogis(q) else pnorm(q)
}

mycdf2 <- function(x,y,rho,logistic=FALSE){
   # *********************************************
   # Calling bivariate CDF: normal or logistic
   # *********************************************
   if(logistic) pbivlogis(x,y,rho) else pbivnorm(x,y,rho,recycle=TRUE)
}


Hope this helps,

Rui Barradas

Às 06:14 de 25/01/21, Steven Yen escreveu:
> Dear All
> 
> Below are calls to functions to calculate bivariate and univariate 
> logistic probabilities.It works for the following sample program (with 
> results p1=p2 and p3=p4), but similar calls in a more elaborated program 
> produced unpredicted results.
> 
> My question is whether I am doing something bad (which I should avoid) 
> in my calls to mycdf2 and mycdf to obtain p2 and p3, respectively. Thank 
> you.
> 
> Steven Yen
> 
> pbivlogis <- function(x,y,rho){
> # *********************************************
> # Bivariate logistic CDF
> # *********************************************
>    p<-(1+exp(-x)+exp(-y)+(1-rho)*exp(-x-y))^(-1)
> return(p)
> }
> 
> mycdf <- function(q,logistic=FALSE){
> # *********************************************
> # Univariate CDF: normal or logistic
> # *********************************************
>    if(!logistic){
>      p<-pnorm(q)
>    } else {
>      p<-plogis(q)
>    }
> return(p)
> }
> 
> mycdf2 <- function(x,y,rho,logistic=FALSE){
> # *********************************************
> # Calling bivariate CDF: normal or logistic
> # *********************************************
>    if(!logistic){
>      p<-pbivnorm(x,y,rho,recycle=T)
>    } else {
>      p<-pbivlogis(x,y,rho)
>    }
> return(p)
> }
> 
> set.seed(123)
> x<-runif(n=5,min=-3,max=3)
> y<-runif(n=5,min=-2,max=4)
> rho<-0.5
> 
> p1<-pbivlogis(x,y,rho); p1
> p2<-mycdf2(x,y,rho,logistic=TRUE); p2
> 
> p3<-mycdf(x,logistic=T); p3
> p4<-plogis(x); p4
> 
> Results
> 
>  > set.seed(123)
>  > x<-runif(n=5,min=-3,max=3)
>  > y<-runif(n=5,min=-2,max=4)
>  > rho<-0.5
>  > p1<-pbivlogis(x,y,rho); p1
> [1] 0.04937376 0.65977865 0.35821101 0.72243120 0.63881214
>  > p2<-mycdf2(x,y,rho,logistic=TRUE); p2
> [1] 0.04937376 0.65977865 0.35821101 0.72243120 0.63881214
>  > p3<-mycdf(x,logistic=T); p3
> [1] 0.2184819 0.8493908 0.3667608 0.9087199 0.9335661
>  > p4<-plogis(x); p4
> [1] 0.2184819 0.8493908 0.3667608 0.9087199 0.9335661
>  >
> 
> ______________________________________________
> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.



More information about the R-help mailing list