Date:         Tue, 23 Oct 2007 14:58:24 -0700
Reply-To:     surendar <surendar.gopal@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         surendar <surendar.gopal@GMAIL.COM>
Organization: http://groups.google.com
Subject:      Re: Time Series for Mortgage Data
Comments: To: sas-l@uga.edu
In-Reply-To:  <685b29a0710231345w1a1b6ce2q87d05d12fd6d3a1e@mail.gmail.com>
Content-Type: text/plain; charset="us-ascii"

On Oct 23, 1:45 pm, samuel.cro...@GMAIL.COM (Samuel Croker) wrote:
> Hi Surendar,
>
> I found that this book was very helpful for me making the transition
> from switching from R to SAS
> (http://support.sas.com/publishing/bbu/companion_site/57275.html) for
> time series analysis and forecasting.  I would recommend that you do
> what Sigurd suggested before diving into the data.  Forecasting things
> that come even partly from human activities can be extremely difficult
> and sometimes it seems that reading chicken entrails might be more
> effective...
>
> Good luck
>
> Sam
>
> On 10/23/07, Sigurd Hermansen <HERMA...@westat.com> wrote:
>
>
>
>
>
> > Surendar:
> > For example, from simple to complex ....
> >http://recenter.tamu.edu/TGrande/vol7-3/1406.html
> >http://www.springerlink.com/content/q072357p32160747/
> >http://goliath.ecnext.com/coms2/summary_0199-3074448_ITM
>
> > S
> > -----Original Message-----
> > From: owner-sa...@listserv.uga.edu [mailto:owner-sa...@listserv.uga.edu]
> > On Behalf Of surendar
> > Sent: Tuesday, October 23, 2007 3:12 PM
> > To: sa...@uga.edu
> > Subject: Re: Time Series for Mortgage Data
>
> > Thank you  Sigurd,Mathew
>
> > Mathew other than these procedures anything else I have to look in
> > to.Like the assumption.
>
> >   Sigurd before I posted this question I searched for literature on
> > Mortgage which was a wild search never found anything and I am still to
> > do on prediction or forecating search.You talked about turning points
> > can you tell me more on that.
>
> > Surendar
>
> > On Oct 23, 11:29 am, HERMA...@WESTAT.COM (Sigurd Hermansen) wrote:
> > > Surendar:
> > > Let me suggest that you first review the literature on forecasting of
> > > market prices. In stock, real estate, and other highly competitive
> > > markets, current prices tend to reflect expectations of future prices.
>
> > > Forecasting models predict better in hindsight than in prospect.
>
> > > The SAS/ETS licensed product has a wealth of documentation on-line.
> > > Once you have a basic understanding of the theoretical issues (serial
> > > correlation, stability, lagged error terms, etc.), introductory
> > > sections will help you understand which procedures would be
> > > appropriate.
>
> > > As many homeowners in the USA will attest today, forecasts based
> > > solely on historical observations can be dangerously misleading. Watch
>
> > > for those turning points. Only the best forecasting models have a
> > > chance of forecasting turning points. S
>
> > > -----Original Message-----
> > > From: owner-sa...@listserv.uga.edu
> > > [mailto:owner-sa...@listserv.uga.edu]
>
> > > On Behalf Of surendar
> > > Sent: Tuesday, October 23, 2007 1:43 PM
> > > To: sa...@uga.edu
> > > Subject: Time Series for Mortgage Data
>
> > > hello,
>
> > >  My dependent variable in Sale Value of Home and my IV variable would
> > > be both continuous and categorical.
>
> > > I cannot use general regression or GLM or mixed since the data is
> > > collected with respect to time.
>
> > > The data is collected since 1974.I want to predict the value of home
> > > for the 2008 say.
>
> > > Can someone suggest a book.Example with SAS and SPSS
>
> > > Thanks- Hide quoted text -
>
> > > - Show quoted text -
>
> --
> Samuel T. Croker
> Lexington, SC & Bethesda, MD- Hide quoted text -
>
> - Show quoted text -

Thanks Samue,Sigurd,Mathew.I would do the ground work first and then
start on the data.Before that I have one other question.

I have worked on survival analysis Proc Genmod.I have a lot and there
are many houses.In a lot for a particular year home is not built but
surrounding homes are built and there is value for them.How will I
deal with data that was dead long time but in future have value have
value now.Can this be survival data.

Thanks once again

Thank you.