Date:         Thu, 4 Nov 2010 13:39:24 -0700
Reply-To:     "Parise, Carol A." <PariseC@sutterhealth.org>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         "Parise, Carol A." <PariseC@sutterhealth.org>
Subject:      Re: Comparing Models
Comments: To: "RAbraham@aaalife.com" <RAbraham@aaalife.com>
In-Reply-To:  <OF7AE6EAA3.A87A4452-ON852577D1.00672C91-852577D1.00681070@aaalife.com>
Content-Type: multipart/alternative;

I would use the regression coefficients to compute the predicted probability of people with different characteristics choosing product X.

For example,  the predicted probability for choosing product X will likely be different for females between 20-30 years versus males between 60-70 years old. You can base which product to mail based on some groups of characteristics.






________________________________
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of <R. Abraham>
Sent: Thursday, November 04, 2010 10:56 AM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Comparing Models


Hi List,
I was wondering if someone could provide me some suggestions.

I want to compare the scores of different logisitic regression (acquisition) models for different products, and based on the scores be able to select which product to mail to a prospect. I know that comparing raw scores would not be accurate. If the scores are normalized for each of the models would the comparison of scores still make sense to select which product offer to mail. Is there any other method?

Pls let me know if it's not clear. Thank you.


Renji Abraham

[text/html]