🎉 🎉 RAPIDMINER 9.10 IS OUT!!! 🎉🎉

Download the latest version helping analytics teams accelerate time-to-value for streaming and IIOT use cases.

CLICK HERE TO DOWNLOAD

Problem With The Apply Threshold Function

SelaTheRockSelaTheRock Member Posts: 2 Newbie
So I tried to make a Logistical Regression on a data and when I run it the apply threshold gets an error saying:
"Predicted label is incompatible"
"The values of the predicted label does not match the model expectation. Value {0} unknown".

Best Answers

  • lionelderkrikorlionelderkrikor Moderator, RapidMiner Certified Analyst, Member Posts: 1,185   Unicorn
    Solution Accepted
    Hi @SelaTheRock,

    In attached file, the working process...
    There was an inconsistency between the setting of classes in Create Threshold operator (Yes, No) and the expected output of the label (true/false).
    Moreover I done several modifications in your process in order to become more relevant.

    To go further, you have an highly imbalanced dataset (true = 7654 / false = 58), that 's why your logistic regression model predicts only
    the class = True.
    I think it's not what you want.
    To improve your process, you can use the Smote Upsampling operator as a preprocessing step in order to increase artificially your minority class(false). by this way you will be able to predict correctly your minority class...

    Good luck ! 

    Regards,

    Lionel
    SelaTheRock
  • SelaTheRockSelaTheRock Member Posts: 2 Newbie
    Solution Accepted
    Hi @lionelderkrikor
    Thank you so much, You helped me a lot I really appreciate the quick answer.

    lionelderkrikor
Sign In or Register to comment.