Date:         Mon, 3 May 1999 16:44:57 -0400
Reply-To:     Mike Davenport <Mike.Davenport@RICHMOND.PPDI.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Mike Davenport <Mike.Davenport@RICHMOND.PPDI.COM>
Subject:      Re: Interpolation imputation for missing
Comments: To: Shiansong Li <shiansong@YAHOO.COM>
Content-Type: text/plain; charset=us-ascii

The formula you want is
 
y_i = ((y_n  -  y_m)*(x_i  -  x_m)/(x_n  - x_m)) + y_m
 
where
      y_i   is the missing value to interpolate
      x_i   is the associated obs for y_i
     y_n  is the non-missing result on the upper end
     x_n   is the associated obs for y_n
     y_m  is the non-missing result on the lower end
     x_m  is the associated obs for y_m
 
Shiansong Li wrote:
 
> Hello, all,
>
> How to write SAS to impute missing values by linear interpolation (not
> locf).
>
> Data itpltn;
> obs value;
> 1   6
> 2   .
> 3   .
> 4   12
> 5   .
> 6   11
>
> The output data set should be like this:
> obs value;
> 1   6
> 2   8
> 3   10
> 4   12
> 5   11.5
> 6   11
>
> Thanks in advance.
>
> _________________________________________________________
> Do You Yahoo!?
> Get your free @yahoo.com address at http://mail.yahoo.com