Options

[SOLVED] Plotting cumulative distributions

tennenrishintennenrishin Member Posts: 177 Contributor II
edited November 2018 in Help
The cumulative distribution is a useful visualization because it does not require arbitrary discretization. What is the easiest way to report a cumulative distribution of a set of real-valued examples in RM?

(Currently I'm doing Sort => GenerateID => Set Role => etc.)

Answers

  • Options
    MariusHelfMariusHelf RapidMiner Certified Expert, Member Posts: 1,869 Unicorn
    Let's do this at the example of a1 of the Iris dataset. Just use the original dataset, go to the Advanced Charts, then:
    - put a1 onto both domain and the numerical axis.
    - Click on the Series: a1
    - set Aggregation to count
    - under Windowing, set "grab left" to -1
    - click the "configure grouping" link, click configure and specify the desired amount of bins

    Please let me know if you have any additional questions.

    Best regards,
    Marius
  • Options
    MariusHelfMariusHelf RapidMiner Certified Expert, Member Posts: 1,869 Unicorn
    Oh, and in addition, if you want the distribution by label, just put the label onto the Color Domain and set the grouping to Distinct Values.

    :)
  • Options
    tennenrishintennenrishin Member Posts: 177 Contributor II
    Thank you!
Sign In or Register to comment.