reasons for getting different results

Haifa_G7Haifa_G7 Member Posts: 1 Newbie
Greetings, 
I'd like to thank you in advance for your help and efforts
I'm newbie to rapid miner so excuse me if my question was too simple 

but I've encountered a problem with using the same dataset and process shared by a friend of mine,
I've not changed anything in the models used or parameters yet I get completely different results from her.
the process contains split validation with decision tree model. 

Thank you.
Tagged:

Answers

  • varunm1varunm1 Moderator, Member Posts: 1,116   Unicorn
    Hello @Haifa_G7

    Can you check if the "local random seed" parameter in split validation operator is set? That might be one reason as test and train data might differ between both of you. If you could post the process here, we can check it. You can attach .rmp file here.
    Regards,
    Varun
    https://www.varunmandalapu.com/

    Be Safe. Follow precautions and Maintain Social Distancing

  • [Deleted User][Deleted User] Posts: 0 Learner III
    edited March 24
    @Haifa_G7

    Hello

    1) Is your detaset balance? 
    2) Do you have any single label in your dataset?
    3) Also for split validation, did you and your friend use the same for train and test part of dataset?

    Regards
    mbs
  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,321  RM Data Scientist
    the reason can be a different random seed, that's why @varunm1 mentiones the random seed. If one of you two are using a very old PC, and thus a 32bit architecture, it may be that you get different random numbers even with the same seed.

    Best,
    Martin

    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
    [Deleted User]
Sign In or Register to comment.