locked
Rows Weights Are Not Working RRS feed

  • Question

  • I'm testing the following experiment:

    Using Row Weights to Penalize False Negatives

    The weights column has no effect to the results, even if I don't use the "New column name" feature of the Edit Metadata module.

    Am I missing anything?

    Thank you.




    Luca Zavarella, Mentor & Technical Director | Data Science MPP | Twitter: @lucazav

    Wednesday, September 26, 2018 7:57 PM

All replies

  • Ciao Luca!

    It seems that there is an old known issue regarding the no effect of introducing a weight column, regardless of whether you change or not the name of the column.

    Maybe the workaround would be to replicate rows for which class = 1, or implement an R or Python script to do it by hand.

    Best!

    Thursday, September 27, 2018 8:09 AM
  • Ciao Ariel!

    I read about the workaround in a post of years ago. I hoped the bug was solved, because I wanted to use weights to try to solve an heteroscedasticity issue in a linear regression.

    @Microsoft guys: any chance the bug will be fixed?

    Thank you


    Luca Zavarella, Mentor & Technical Director | Data Science MPP | Twitter: @lucazav

    Thursday, September 27, 2018 12:19 PM
  • Thanks for your feedback here. I will forward it to the product team and get back to you once I get some response. Thanks again for bringing the issue to us.

    Regards,

    Yutong

    Monday, October 1, 2018 3:33 PM