Re: linear interpolation through missing data

[Sundar Dorai-Raj <sundar.dorai-raj@PDF.COM>]

Lambert.Winnie wrote:

> S-PLUS6 on Windows XP
>  
> I found this question in 3 emails in the archive, but did not see an 
> answer.  I have looked in my S-PLUS books, but perhaps am looking in the 
> wrong place for the wrong thing.  I want to do a relatively simple 
> linear interpolation across missing values in a vector, e.g.  ...850 862 
> 875 NA NA 927 946... weighted by another non-missing (and never 
> missing!) variable.  Is there a function for this?  If not, I can trudge 
> through and figure out an algorithm, but would rather not waste time if 
> I can do it in a one-liner.
>  
> Thanks much.
>  
Try ?approx. You'll need to supply an `x' but here's an example to help:

set.seed(1)
n <- 100
y <- sort(rnorm(n))
x <- seq(n)
na <- sample(n, 3) # remove 3 points
guess <- approx(x[-na], y[-na], xout = x[na])$y
cbind(true = y[na], guess)

--sundar