The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here

Question data

andre5007andre5007 Member Posts: 22 Learner II
edited May 2021 in Help 
I have these two csv, in which both csv have several feats.
Feat1- model, Feat2-power measure, Feat3- is something that this object has or does not have, being 1 has and 0 does not, Feat4 is a feature that I don’t know what it is, Feat5- device installation date, Feat6 / 7- It is the latitude and longitude and feat 8 is the number
maintenance interventions.
In the CSV Training I have values ​​for feat 8 and in the Test no.
My goal is to estimate the Feat 8 for the Test set.
How can I do this? 
Thanks

Best Answers

  • yyhuangyyhuang Administrator, Employee-RapidMiner, RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 364 RM Data Scientist
    Solution Accepted
    Hi @andre5007, it looks your prediction target is numerical (integers). Are you sure you want to build decision tree or any predictive model for classification, rather than regressions? I would parse the label into numbers and try the regression decision trees or GLM/GBT for regression.
  • andre5007andre5007 Member Posts: 22 Learner II
    Solution Accepted
    Hi @yyhuang
    Why do you think regression decision trees or GLM/GBT for regression is better?
    Thanks
    André
  • yyhuangyyhuang Administrator, Employee-RapidMiner, RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 364 RM Data Scientist
    Solution Accepted
    Hi @andre5007, my point was regression is better than classification here as the model for your data. Because the label is integer. For the difference between regression and classification, https://machinelearningmastery.com/classification-versus-regression-in-machine-learning/

Answers

  • Telcontar120Telcontar120 RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,635 Unicorn
    You should review the RapidMiner tutorials for Cross Validation and for Apply Model. Basically you are going to define Feat 8 as the label and build your model on that, and then you are going to save that model and apply it to the 2nd dataset.
    Brian T.
    Lindon Ventures 
    Data Science Consulting from Certified RapidMiner Experts
  • andre5007andre5007 Member Posts: 22 Learner II 
    Ok, I will try to see and do, if you have any questions then can you help me?
  • andre5007andre5007 Member Posts: 22 Learner II
    Can someone tell me if I'm going in a good way please?


  • yyhuangyyhuang Administrator, Employee-RapidMiner, RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 364 RM Data Scientist
    Hi @andre5007,

    The workflow looks fine if you have your own test set. However, as Brian mentioned above, cross validation is always a smart option on your training set.

    https://academy.rapidminer.com/learn/article/cross-validation
    https://academy.rapidminer.com/learn/video/validating-a-model
    https://rapidminer.com/blog/validate-models-cross-validation/

    HTH!

    YY

  • andre5007andre5007 Member Posts: 22 Learner II
    Now I noticed that I was wrong on the print I sent, because it was not the one I wanted to have selected.

    I put a filter at the beginning because it had a value that was missing and because of that it gave an error.

    Then in the cross validation, I placed the decision tree inside the process at the training site and in the test the apply model and performance.

    Then I linked the cross validation to another apply model and in that apply model I also put the test data set where I have to define feat 8.

    Do you think you should change anything in the operators parameters? Because I didn't change anything just when it was necessary to be able to run the process.

    What do you think I can improve? Or if I am now on the right path? 

    Thanks
    Best regards
    André


  • Telcontar120Telcontar120 RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,635 Unicorn
    Looks like a good setup for basic model construction and validation with an additional out-of-sample validation. 
    Brian T.
    Lindon Ventures 
    Data Science Consulting from Certified RapidMiner Experts
  • andre5007andre5007 Member Posts: 22 Learner II 
    Can you explain how I can do to improve the value that I mark in red?
Thanks
  • rugmanasokanrugmanasokan Member Posts: 1 Learner I
    As a model for your data, regression is better than classification. Due to the integer nature of the label. In order to understand the difference between regression and classification - https://nimblebox.ai/blog/regression-machine-learning
Sign In or Register to comment.