Is there multi-dimensional scaling support in rapidminer?

it_mjjiang

or in rapidanalytics or rapidnet?

Thanks!

IngoRM

Hi,

what is multi-dimensional scaling? Can you explain what you want to achieve? If it is that what I assume the answer is: yes.

Cheers,
Ingo

Preko

This was also bugging me for a while, tried to solve it in RapidMiner and I think the answer is no.
By multidimensional scaling (MDS) I mean this: http://en.wikipedia.org/wiki/Multidimensional_scaling

For example, here is the Matlab version:
http://www.mathworks.com/help/toolbox/stats/mdscale.html
It has a distance matrix as an input and it maps all the points usually to 2D/3D.

It is an old technique, but there is still active research on it. A unified solution was presented at KDD 2010:
http://portal.acm.org/citation.cfm?id=1835948

I think it would be a great addition to RapidMiner.