ANNOUNCEMENT: RAPIDMINER 9.1 BETA HAS BEEN RELEASED TODAY!   PLEASE DOWNLOAD AND GIVE FEEDBACK. ENJOY AND HAPPY RAPIDMINING!   -- @sgenzer – Community Manager

Discriminant Analysis

smajumdarsmajumdar Member Posts: 2 Contributor I
edited November 9 in Help

I wish to understand the output (from model port) of both LDA & QDA (Linear and Quadratic Discriminant Analysis) operators. When running on a small dataset (1 binomial label & 2 real number predictors), I'd expect some kind of coefficients but I get only apriori probabilities of the 2 groups represented by the binomial label. Am I missing something?

QDA.png 40.2K
Tagged:

Answers

  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member Posts: 196   Unicorn
    Hey @smajumdar, do you mind to share your XML process with us? Both LDA and QDA have two ports, and you have only one connected which is the mod port. Perhaps you want to see your data? For that there is the exa port but I don't see that in your picture.
  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member Posts: 196   Unicorn
    Duh me! I forgot something too! Here is a full example of applying Quadratic Discriminant Analysis on my favourite dataset: the Titanic.



    You are getting the model, but to get real insights, you should apply that model to some more data. In my case, once I apply the model on my unlabeled data, I get the "Survived" column as a prediction based in QDA.



    Please find the example attached.

    Hopefully this helps you solving your issues. Enjoy!

    Rodrigo.
    mschmitzsmajumdar
  • smajumdarsmajumdar Member Posts: 2 Contributor I
    Thanks buddy
    rfuentealba
Sign In or Register to comment.