Date: Fri, 22 Mar 1996 17:42:48 GMT
Reply-To: cepotter <cepotter@CCS.CARLETON.CA>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: cepotter <cepotter@CCS.CARLETON.CA>
Organization: Carleton University
Subject: Re: Validating a multiple regression model
cepotter (email@example.com) wrote:
> Does any one know how to get SAS to generate predicted values and
> associated confidence intervals for "future observations"?
> I am a biology graduate student at Carleton U, working on my Masters
> Thesis in ecology. I have a 5 term regression model generated from 100
> observations. I have 50 other observations which I held back, when
> constructing the model. I now need to determine the validating of the model
> through its ability to predict the remaining 50 observations. I can plug
> the parameter estimate co-efficients with the predictor values for the 50
> observations into an excell spreadsheet and come up with the predicted
> values but not confidence intervals for those predicted values to see if
> the cooresponding observed values fit into the confidence range.
> All the other students in my lab are stumped, we can find complex
> equations in regression text books, but cannot find the programming
> statements in the SAS manual----perhaps we do not know the proper "buzz
> words" or are not being create enough in navigating the index.
> HELP HELP HELP