Menu

MDS of data in rkward

Help
2007-10-08
2013-01-15
  • Sebastian Boell

    Sebastian Boell - 2007-10-08

    I have data representing the similarity between ten different databases in a matrix format. Now I like to create a similarity map using multi dimensional scaling (MDS). Can somebody tell me if I can do such a thing using rkward? I could not find a self expanatory entry in the menues and searching through the formum, wiki and even the whole net for rkward and mds was unsuccesful. Appreciate any help.

     
    • P Kapat

      P Kapat - 2007-10-08

      Hi elevel,

      My understanding is that we do not have any existing plugins for your purpose. But, you could always use the Rconsole from RKWard, if you know the right functions in R.

      Sorry, not being "objectively" helpful!
      Regards,
      PK

       
    • Thomas Friedrichsmeier

      Hi,

      as PK has pointed out, RKWard does not have a GUI for MDS, yet. And it's not the only thing you'll find missing in the menus. But RKWard is based on a very powerful statisticas system (R) and you can access all that functionality using the RConsole. A useful starting point is the "Help Search" (at the bottom). Searching for MDS, there, will point you to functions cmdscale() in package stats, and isoMDS() in package MASS.

      Juding by your name, you are likely German, so this more hands-on instruction may be of interest to you: http://www.wiwi.uni-bielefeld.de/~frohn/Mitarbeiter/Handl/mvamdsr.html .

      Regards
      Thomas

       
    • Sebastian Boell

      Sebastian Boell - 2007-10-10

      Hi PK and Thomas,

      thank you both very much for your replies in that matter. It is good to know that R supports MDS and therefore that there is a potential to include this capability into the GUI of RKward in future. Unfortunately I am a newbee to both R and MDS therefore I decided to deal with one at the time, which will be MDS at this stage.

      Also very much for the link to the description on how to do MDS in R, and yes I am German. Thank you very much again for your replies.

      Regards
      Sebastian

       
MongoDB Logo MongoDB