RapidMiner 9.8 Beta is now available

Be one of the first to get your hands on the new features. More details and downloads here:

GET RAPIDMINER 9.8 BETA

Logistic Regression_AUC_Store and retreive model

SA_HSA_H Member Posts: 26 Contributor II
edited December 2019 in Help
For logistic regression classifier, is it possible to:

1- Draw AUC for both training and testing data "in the same model" without using cross validation operator "I mean use only split data operator".
2- How to save model and use it to train new dataset.

I would appreciate if the answer can be associated with a figure or video or process for better illustrating.

Answers

  • lionelderkrikorlionelderkrikor Moderator, RapidMiner Certified Analyst, Member Posts: 1,085   Unicorn
    Hi @SA_H,


    1- Draw AUC for both training and testing data "in the same model" without using cross validation operator "I mean use only split data operator".

    Please find in attached file , a process which calculates the ROC curves for both training set and test set.
    Please note that the process will create 2 figures (one for the training part, one for the test part), so the 2 curves are not on the same figure.


     2- How to save model and use it to train new dataset.


    We do not train a "new dataset", we train a model.
    So, you mean how to save model and use it to score a new dataset ?
    if so, you can use the Store operator to save the model, and the Retrieve operator to recover your trained model and then you can score a new dataset by using the Apply Model operator.

    Hope this helps,

    Regards,

    Lionel

    Tghadially
  • SA_HSA_H Member Posts: 26 Contributor II
    edited December 2019
    Thank you Lionel. I would appreciate  if you can provide an image for retrieve saved model to test new dataset.
    Another inquiry if possible, how can I remove ROC-threshold and keep only ROC in the figure.
    Tghadially
Sign In or Register to comment.