Support Vector Regression not handling nominal attributes

islem_h
islem_h New Altair Community Member
edited November 5 in Community Q&A
Hello everyone :) 

I have a mostly categorical data except for some attributes icluding the target variable (numerical). I a trying to apply Support Vector Machine on it for a prediction task.
It would be too easy to just apply the operator Nominal to Numerical directly.
I tried the Auto Model to get inspiration from it, but I couldn't really follow what it was doing.

A little help in this would be much appreciated.
Thank you in advance!

Answers

  • MartinLiebig
    MartinLiebig
    Altair Employee
    Ehm, why don't you use Nominal to Numerical?
  • islem_h
    islem_h New Altair Community Member
    It gives back high values of RMSE in comparison to the Auto Model 
    and the difference is only in the way of dealing with the nominal attributes.
  • MartinLiebig
    MartinLiebig
    Altair Employee
    Hi Islem,

    did you use the evolutionary feature selection option in AM?

    BR,
    MArtin 
  • islem_h
    islem_h New Altair Community Member
    I did, and I also used the same selected attributes in my own process
  • Telcontar120
    Telcontar120 New Altair Community Member
    Nominal to Numerical is definitely the way to go, and the dummy coding method should be appropriate for SVM.  Did you use integer coding by chance instead?  That is usually not the correct way to do it if the underlying nominals are not ordinal/scalar.  Maybe you could post process or data samples for further diagnosis.

  • islem_h
    islem_h New Altair Community Member
    Here is a screenshot of the process or is there another way to share it?
    PS: I used the dummy coding method already. 
  • lionelderkrikor
    lionelderkrikor New Altair Community Member
    Hi @islem_h,

    To share your process (in order we can execute it) : 

    Note: This solution requires the "XML" panel which can be opened in the "View" menu and then "Show Panel".  Activate the XML panel if you did not do this before.

    Open your process in RapidMiner and open the XML panel. If you can't find it, make sure to follow the note above.

    Copy the XML code from there and paste it somewhere else, for example into a forum post here on the community portal.  By the way, if you post your XML here, please use the code environment which you get by clicking on the icon Code  in the toolbar of the post.


    Can you share your dataset(s) too ?

    Regards,

    Lionel