🎉 🎉 RAPIDMINER 9.10 IS OUT!!! 🎉🎉

Download the latest version helping analytics teams accelerate time-to-value for streaming and IIOT use cases.


"Numeric Dimension Reduction (PCA, etc) for Cluster Analysis"

WI_NobleWI_Noble Member Posts: 5 Contributor II
edited May 2019 in Help
I'm working on a clustering application and want to reduce the numeric attributes, especially correlated ones, prior to clustering.  A proven technique to doing so is PCA, which RapidMiner contains.  What I also want to do is the Varimax rotation algorithm to "load" a minimum number of attributes per component to improve interpretability.  In a prior post, it was mentioned that PCA rotation (Varimax, etc.) is on the future wish list.  I have the following questions on this:

1.) How close is RapidMiner to including rotation in future releases?
2.) If not "close", would I be better off looking at using an R package within RapidMiner to accomplish this?
3.) What other techniques have people used in RM to reduce numeric dimensions to a minimum, non-correlated set that are just as good/better than PCA for cluster analysis?

Sign In or Register to comment.