Need one regression model or multiple?

jawon — Thu, 22 Jul 2010 21:34:23 GMT

This is more a general regression question than specific to sas procs...

Let's say I'm trying to create a model that predicts how much my customers are likely to spend at my store. But looking at my existing customers, I have a hunch that customers who have bought product X are different than those who have bought product Y.

In other words, if I were to build a model from product X customers, the final variables and coefficients could be very different from the model built from product Y customers.

Do I have to build a separate model for each customer segment and somehow blend the results? Or is there a way to build one model that somehow factors in the differences?

Re: Need one regression model or multiple?

Doc_Duke — Fri, 23 Jul 2010 02:40:42 GMT

In the classic regression sense, you can do this with interaction terms in one model.

As a concrete example, say that X=baby diapers and Y=Viagra. So, you'd expect X-people to be mostly younger and female and the Y-people to be mostly older and male. If X and Y are 0/1 coded, a model might look something like

$$spent = age sex x y age*x age*y sex*x sex*y

Doc Muhlbaier
Duke

topic Need one regression model or multiple? in Statistical Procedures

Need one regression model or multiple?

Re: Need one regression model or multiple?