Loading ...
Sorry, an error occurred while loading the content.

[ai-geostats] Trend surface analysis

Expand Messages
  • Chunhua Zhang
    Hello list, I want to know how to conduct trend surface analysis in SAS. I tried regression model and it seems it is not the right model. Will GLM work? Many
    Message 1 of 3 , Mar 6 11:53 AM
    • 0 Attachment
      Hello list,

      I want to know how to conduct trend surface analysis in SAS.
      I tried regression model and it seems it is not the right model.
      Will GLM work?

      Many thanks!

      Chunhua
    • Mahdi Osman
      I have not used SAS for a while thanks to R, but I know that R s spatial library will do a good job in trend surface analysis. If you are using R, you could
      Message 2 of 3 , Mar 6 3:41 PM
      • 0 Attachment
        I have not used SAS for a while thanks to R, but I know that R's "spatial"
        library will do a good job in trend surface analysis.

        If you are using R, you could load the spatial library and check it out.

        >> library(spatial)
        >> help(package = "spatial")

        ?spatial will give you some helpful information. There is a documentation on
        spatial library that you can find on the web.

        If you need some examples on that please contact me offline.

        Cheers
        Mahdi


        > Hello list,
        >
        > I want to know how to conduct trend surface analysis in SAS.
        > I tried regression model and it seems it is not the right model.
        > Will GLM work?
        >
        > Many thanks!
        >
        > Chunhua
        >
        >

        --
        -----------------------------------
        Mahdi Osman (PhD)
        E-mail: m_osm@...
        -----------------------------------

        SMS bei wichtigen e-mails und Ihre Gedanken sind frei ...
        Alle Infos zur SMS-Benachrichtigung: http://www.gmx.net/de/go/sms
      • Roger Bivand
        ... One possible reason for problems with trend surface and standard OLS functions is that if, say, coordinates are in UTM metres, y^3 in particular will
        Message 3 of 3 , Mar 7 1:17 AM
        • 0 Attachment
          On Mon, 7 Mar 2005, Mahdi Osman wrote:

          > I have not used SAS for a while thanks to R, but I know that R's "spatial"
          > library will do a good job in trend surface analysis.
          >
          > If you are using R, you could load the spatial library and check it out.
          >
          > >> library(spatial)
          > >> help(package = "spatial")
          >
          > ?spatial will give you some helpful information. There is a documentation on
          > spatial library that you can find on the web.

          One possible reason for problems with trend surface and standard OLS
          functions is that if, say, coordinates are in UTM metres, y^3 in
          particular will become a big number for anywhere not close to the Equator,
          and so on. In software suited to trend surface, the coordinates are
          typically rescaled to lie between -1 and +1 - this is what Prof. Ripley
          does in surf.ls() in the spatial package for S (runs on both S-Plus and
          R), and what Edzer Pebesma does in gstat (both for S and freestanding)
          among others. My guess would be that trend surface will work in SAS too if
          the coordinates are rescaled before fitting. New data for prediction will
          need to be rescaled in the same way.

          >
          > If you need some examples on that please contact me offline.
          >
          > Cheers
          > Mahdi
          >
          >
          > > Hello list,
          > >
          > > I want to know how to conduct trend surface analysis in SAS.
          > > I tried regression model and it seems it is not the right model.
          > > Will GLM work?
          > >
          > > Many thanks!
          > >
          > > Chunhua
          > >
          > >
          >
          >

          --
          Roger Bivand
          Economic Geography Section, Department of Economics, Norwegian School of
          Economics and Business Administration, Breiviksveien 40, N-5045 Bergen,
          Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93
          e-mail: Roger.Bivand@...
        Your message has been successfully submitted and would be delivered to recipients shortly.