Can 10-fold cross-validation be implemented in Rapidminer Automodel?

515bm1003515bm1003 Member Posts: 3 Contributor I
edited March 2020 in Help
Can 10-fold cross-validation be implemented in Rapidminer Automodel? By default, it is implementing 2-fold cross-validation. I need to do 10-fold cross-validation for a research work.
Thanking you in Advance.

Regards,
Suraj

Answers

  • varunm1varunm1 Moderator, Member Posts: 1,207 Unicorn
    Hello @515bm1003

    Actually, auto model doesn't implement any cross validation. It does 60:40 (train :test) split validation with the 40% test data divided into 7 sets and tested independently. The performance of these 7 sets of test data is aggregated.

    You need to open the automodel process and manually change it to cross validation. It is doable but a bit complex as you need to shift lot of operators into cross validation and delete some. You can try it by opening the process
    Regards,
    Varun
    https://www.varunmandalapu.com/

    Be Safe. Follow precautions and Maintain Social Distancing

  • [Deleted User][Deleted User] Posts: 0 Learner III
    edited March 2020
    @515bm1003

    Hello
    You can easily use cross validation operator and make your process in studio. Also you can use all algorithms that you see in Auto model. It is easier than adding cross validation to Auto model

    Regards
    mbs
  • Telcontar120Telcontar120 Moderator, RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,635 Unicorn
    You can also take the process created by AutoModel and swap out the split validation (the default setup as Varun described) for a cross-validation instead.  This is not that complicated although it does require you learn how to use the process design view and not rely entirely on the AutoModel wizard.
    Brian T.
    Lindon Ventures 
    Data Science Consulting from Certified RapidMiner Experts
  • YinYin Member Posts: 17 Contributor II
    edited January 2023
    Hi @Telcontar120@varunm1, You both mention that one should replace split validation with cross-validation, and I just posted a question about this here because it seems confusing to me whether auto model is actually using cross-validation or not (See link below). In the link below i thought that auto model is actually using CV, but here i see that it is not. I have not seen anywhere on the training website OR here, a comprehensive way to swamp split validation with CV. Since everyone recommends CV, especially @IngoRM 
    why isnt the default CV? Or can someone please post a helpful tutorial to change split validation to CV?

    Edit** also this (https://community.rapidminer.com/discussion/55112/cross-validation-and-its-outputs-in-rm-studio) was a great read but was not intended for implementing Cv in auto model. 

    Thank in advance.

    https://community.rapidminer.com/discussion/59982/is-cross-validation-automatically-implemented-in-auto-model#latest
Sign In or Register to comment.