LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous (more recent) messageNext (less recent) messagePrevious (more recent) in topicNext (less recent) in topicPrevious (more recent) by same authorNext (less recent) by same authorPrevious page (February 2007, week 4)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Thu, 22 Feb 2007 23:08:03 -0800
Reply-To:     David L Cassell <davidlcassell@MSN.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         David L Cassell <davidlcassell@MSN.COM>
Subject:      Re: Baseline Calculation in Proc Phreg
In-Reply-To:  <200702230456.l1N2cosO002659@mailgw.cc.uga.edu>
Content-Type: text/plain; format=flowed

ppatrawala@INDUCTIS.COM wrote: > >Hi > >I am building a survival model to calculate customer lifetime. >My model includes time dependent covariates (that is variables that take >different values at different points in time - e.g. salary, spend etc). > >I am using the procedure Phreg (SAS ver 9) and have been unable to >calculate the baseline survivor function for time dependent data.(Since SAS >does not give a baseline for time dependent data) > >What I can understand is that we can only find the point estimates and not >the baseline survivor function for the above. > >I have come across several SAS SUGI papers, and referred the book on >Survival analysis by Paul Allison. However, all of them suggest that the >calculation of baseline function is not possible for time varying >covariates. I have also incorporated the counting style process of input >but i am unable to calculate the baseline intercept. > >How do I go about trying to implement a model of such a nature. Is there a >way to do so in Proc Phreg or an alternate solution to using Proc Phreg. > >Thanks >Perseus

First, I am not yet convinced that you have a survival analysis problem. Can you tell us more about your data and your data sources and your meta-data? What *exactly* are you trying to build a predictive model for?

Then I am not convinced that PHREG would be the way to go anyway. I rather doubt that your data are going to meet the proportional hazards assumption. So the proc (and the method) would be the wrong thing to do. Have you thought about investigating this issue first?

HTH, David -- David L. Cassell mathematical statistician Design Pathways 3115 NW Norwood Pl. Corvallis OR 97330

_________________________________________________________________ The average US Credit Score is 675. The cost to see yours: $0 by Experian. http://www.freecreditreport.com/pm/default.aspx?sc=660600&bcd=EMAILFOOTERAVERAGE


Back to: Top of message | Previous page | Main SAS-L page