R help - Sep 2011 - UNSOLVED: Fwd: generate correlated qualitative data

The problem remain unsolved. If you have any idea please do suggest
....................

thank you;

Ram H

On Wed, Aug 31, 2011 at 12:06 PM, Ram H. Sharma
<sharma.ram.h@gmail.com>wrote:
> Dear R experts:
>
> I have following problem:
>
> # myfunction
> mfun1 <- function(x) {
>    if (  x == 2){
> xv <- sample(c(2,1,0),100, replace = T, prob = c(0.6, 0.2, 0.2))
> }
> if ( x == 1){
>  xv <- sample(c(1,0),100, replace = T, prob = c(0.6, 0.4))
>   }
> if ( x == 0) {
>  xv <- sample(c(0,0),100, replace = T, prob = c(0.5, 0.5))
>  }
>  return(xv)
>  }
>
> # applying the function
> x = mfun1(1)
>
> # just what I know so far, correlated sample for a quantative data (i do
> not want !)
> library(ecodist)
> y <- corgen(x=x, r=.5)$y
> What I want is to generate new variable that is correlated at the same new
> variable should be a qualitative one with just value 2,1,0 (no decimals).
>
> I have vector of length 100, just an example
> vc <- sample(c(2,1,0),100, replace = T, prob = c(0.6, 0.2, 0.2))
>
> I want to apply the above function to this vector, while holding the
> correlation for correspondind variables - i.e. X1 and X2 has correlation of
> 0.5, X2 and X3 has also correlation of 0.5 and X3 and X4 has correlation of
> 0.5 and so on.
>
> I have already spend considerable time on it, that is how I got to the
> group, sorry for that.
>
>
> --
>
> Ram H
>
>

-- 

Ram H

	[[alternative HTML version deleted]]