FW/BW Selection with predefined Attribute set?

Fred12Fred12 Member Posts: 344 Unicorn
edited November 2018 in Help

hi,

is it somehow possible to start the Forward /Backward Selection algorithm with a starting set of attributes? e.g I not want to begin with all or non of the attributes, but start from a selection of e.g 10 attributes that I chose before..

Best Answer

  • bhupendra_patilbhupendra_patil Administrator, Employee, Member Posts: 168 RM Data Scientist
    Solution Accepted

    In the example if you add a break point before SVM learner operator , you will notice it trains on a1,a2 and other other columns, based on selection.

Answers

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,453 RM Data Scientist
    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
  • Fred12Fred12 Member Posts: 344 Unicorn

    I thought this process just shows how to ignore attributes, the first 2 (a1,a2) are ignored when doing fw-selection, and later just added to the found attributes from fw-selection, or will the fw-selection take into account the a1 and a2 attributes and start from there? I dont think so...

    I put breakpoints at x-validation, and it shows a3, a4 attribute and adds only a1 and a2 at the end of the process..

  • Fred12Fred12 Member Posts: 344 Unicorn

    ok you are right,

    its just weird that the feature names lag behind the process.. e.g for the corresponding row of squared error 154 the attribute is a3.. however this resulted from testing a4... and so on

    Unbenannt.PNG

Sign In or Register to comment.