Remap Binomials - positive class change doesn't work

varvara_karpovavarvara_karpova Member Posts: 1 Contributor I
edited November 2018 in Help

Hi there!

I would need some help with Remap Binomials. For some reason in my processes (please see a crafted example of such a process attached below) this operator doesn't change the class to "positive" as I specify. In essence, my process is build as follows:

- it has an unbalanced dataset, 

- on which I do a 10 times Cross-Validation (using Decision tree), 

- and I need to perform an Oversampling subprocess *within* the Cross Validation.


I noticed that removing Oversampling doesn't cause this problem with Remap Binomials. However, I abslutely need to use Oversampling because my datasets are unbalanced..

I have studied all previous threads on the RM community about Remap Binomials, but unfortunately, they don't help me to solve the issue.


Many thanks for your ideas!


  • Options
    MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,525 RM Data Scientist

    Dear Vavara,


    i have investigated zour problem. You simply need to add the remap binominals operator after your append operator in the upsampling and everything is fine.


    A bit of background:

    RapidMiner uses a Nominal to Integer Mapping. So all your values are mapped to integers, e.g.

    Yes -> 0

    No -> 1

    The Remap Binominals is changing this mapping in a way that the negative class is 0.

    If you append two datasets it is very likely that the two mapping tables are not compatible (i.e in the first one its yes->0 in the second it's none->0). Thus we need to do some remapping. That happens in your append. Unluckily this messes up the postive and negative classes. But forcing it into line after appending solves it.


    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
Sign In or Register to comment.