Visualization of Self-Organizing Maps (U-matrix with points)

adweadwe Member Posts: 2 Contributor I
edited October 2019 in Help

Hi!

I am in doubt of what I'm actually visualizing with my U-matrix of the SOM.

I have a lot of data of which I've removed a label from.
This data is fed to the SOM which are then trained. The output of the SOM are then joined with the removed label from the original data.

This joined dataset is visualized with a U-Matrix - on top of that, the points corresponding to the label is plotted on top of this U-Matrix.
What am I looking at here? I'm in doubt of why and how the points distribute as they do.

I hope it makes sense 

Cheers!

Answers

  • sgenzersgenzer Administrator, Moderator, Employee, RapidMiner Certified Analyst, Community Manager, Member, University Professor, PM Moderator Posts: 2,959 Community Manager

    hello @adwe welcome to the community. Perhaps you can help by posting your XML and data set as described https://community.rapidminer.com/t5/Getting-Started-Knowledge-Base/READ-ME-HOW-TO-POST-QUESTIONS-ON-THE-RAPIDMINER-USER-COMMUNITY/ta-p/49485?

     

    Scott

     

  • adweadwe Member Posts: 2 Contributor I

    Thanks for your answer. 
    Unfortunately I'm no longer in possesion of the XML - my computer broke down..
    But, this is the resulting U-Matrix
    SOM015.pngU-Matrix

    The points are supposed to be the plotted class labels that were joined with the resulting SOM data - but I can't figure why there are so few and how it is supposed to be interpreted.
    I know the basics about how a self-organizing map works and what a U-Matrix shows, but can't seem to understand this.

Sign In or Register to comment.