"Error=Order of attributes is not equal for training and application example set"

wotsiznamizwotsiznamiz Member Posts: 9 Contributor II
edited May 2019 in Help
I am running a text-mining model...  I have two different text data sets -- one for building the model and one for validating.

When I attempt to load the model and apply it to my validation test set, I notice the following error message...

[Warning] Kernel Model: The order of attributes is not equal for the training and the application example set. This might lead to problems for some models.

Does anyone have any idea why I'm getting this error message?
Tagged:

Answers

  • mihaimihai Member Posts: 3 Contributor I
    rookie opinion:
    first, i never got this message in rapindminer so i m just going to give you a general advice that i follow.
    it is a warning so i would ignore it. before ignoring it completely, i would manually check the datasets to be sure everything is ok.
    in the case of those kind of messages i would first check if the warning comes from a difference in the way the datasets are written, i would check to be sure the label is correctly chosen in both datasets  and if i m still unsure, i would try to use a "leave one out" or k cross fold validation, and if the performance is the same as using the test dataset, then everything is ok.
    but generally i ignore the warnings.

  • IngoRMIngoRM Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, Community Manager, RMResearcher, Member, University Professor Posts: 1,751 RM Founder
    Hi,

    in general this is right: most warnings are not errors and can often be ignored. Here I would like to add that in the case of text mining it is important to save the wordlist during training and re-use during model application. Did you do that?

    Cheers,
    Ingo
Sign In or Register to comment.