Toni Zhang wrote:
> I have a project of Discriminant Function Analysis to find items that
> predict group membership to the best possible accuracy for segmentation
> I know the process it to get each respondentís probability score of
> membership into each of the, say 11 segments. There are 6 predict variables
> that used, they are Q1-Q6. Afterwards, I need to multiply the response on
> each item by the coefficients and sum all of the scores and add the
> constant. I then need to compare the 11 scores within each respondent. The
> respondent is placed in the segment that has the biggest score.
> I wonder if there is a good SAS code do this all at once. If there is not,
> what should I do to save the first SAS output and then do the calculation
> and then compare the 11 scores. If you have created and like some beautiful
> SAS code, please share with me if you donít mind.
Seems to me that PROC DISCRIM does exactly this.
Eastman Kodak Company
"It's nothing until I call it!" -- Bill Klem, NL Umpire
"Those black-eyed peas tasted all right to me" -- Dixie Chicks