tcm5026tcm5026 Member Posts: 5 Contributor II
edited November 2018 in Help

I am trying to select the ten most "important" variables out of 24.  The goal is to use the first component of PCA analysis to find these ten variables.  I have set "generations_without_improval" in the WeightGuidedFeatureSelection to 14 - so that ten variables are selection.  I've tried various numbers and sometimes I get the following error and sometimes I do not.  The error I keep getting is, "Process Failed  IndexOutOfBoundException caught: Index:0, Size:0.

My process looks like this,

Does anyone know what this error means?  I've tried searching the documentation but cannot find it.  Does anyone know possibly a different way to use PCA to select the ten most "important" variables (i.e. eliminating the redundant members)?  Thanks for any help as it is greatly appreciated!!!


  • Options
    landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    this error message denotes an internal memory addressing issue, that is simply a bug and not a hand crafted error message standing anywhere in the documentation.
    I would suggest to update to RapidMiner 5.0, since we don't maintain the RapidMiner 4.6 community version anymore. There's a fair chance this issue has already be solved with RapidMiner 5.

Sign In or Register to comment.