Apply the discretize operator at each year

V23
V23 New Altair Community Member
edited November 5 in Community Q&A

Hi RapidMiner community,

I am working on a project where I have to analyze the world economy by calculating the GDP and class them into 4 quartiles at each year.
I’ve used the discretize by frequency operator. However, when I apply it, it discretizes the whole dataset into 4 quartiles (that contains data for more than 60 years), and I would like to apply this operator and find the 4 quartiles at each year (so it will have different countries within the first quartile in 1950 and in 2000 for example).

Do you have any ideas how I could proceed?

Thanks for your time!

Tagged:

Best Answer

  • MartinLiebig
    MartinLiebig
    Altair Employee
    Answer ✓

    Hi,

     

    the easiest way is to use Group into Collection from operator toolbox, group into years and use Loop Collection afterwards to calculate yourquantiles per year.

     

    Cheers,

    Martin

Answers

  • MartinLiebig
    MartinLiebig
    Altair Employee
    Answer ✓

    Hi,

     

    the easiest way is to use Group into Collection from operator toolbox, group into years and use Loop Collection afterwards to calculate yourquantiles per year.

     

    Cheers,

    Martin

  • V23
    V23 New Altair Community Member

    Great that works! I hadn't those operator and had to download the extension :).
    However, after the "loop collection" we have to put an specific output. The best I found is the "Select" operator, but I can select only one year at the time. Is there a solution where I can select more than one, or even better have the whole data treated in the loop for each year on the same table?

    Thanks a lot for your previous answers, that help me a lot!

  • MartinLiebig
    MartinLiebig
    Altair Employee

    Hi,

     

    have a look at the Append operator. it can make a table out of the collection again.

     

    Cheers,

    Martin

  • V23
    V23 New Altair Community Member
    That's exactly what I needed!
    Thanks a lot, I've been turning around for quite a while :)