none
Visual interface - Evaluation after classification RRS feed

  • Question

  • I built a model to investigate the differences between two multiclass classification algorithms, and the visualization of the evaluation looked very surprising: showing only a few values in two rows. Am I using it wrong?

    When I was doing the same with Azure ML Studio, I could get the accuracy and the confusion matrix as well, the two algorithms were shown next to each other.

    The first issue I have with the new version is that I'm not sure which row represents which algorithm (though I guess first row represents the algorithm on the right side, or the one that finished training first). On the other hand, confusion matrix was providing a very good indication on how to improve the model and the features.

    Will these be fixed, at some point, or do I only have to change something? Will there also be more information provided than we had in Azure ML Studio? :)

    Thank you for all the info! :)


    • Edited by Eva Pardi Wednesday, October 16, 2019 5:48 PM
    Wednesday, October 16, 2019 5:47 PM

Answers

  • Hi Eva,

    Thanks for reaching out here. All functions of Azure Machine Learning Studio will be lighted up in Azure machine learning service in the near future. And also thanks for your feedback for Visual interface, we will investigate it internally and do the necessary change and update.

    Regards,

    Yutong

    • Marked as answer by Eva Pardi Thursday, October 17, 2019 6:09 AM
    Wednesday, October 16, 2019 6:13 PM
    Moderator