Aleks Jakulin schrieb: ... > Yes, by using variable clustering, for example. I noticed that there are > procedures in SAS, and in Frank Harrell's R/S libraries for this purpose. Has > anyone compared PCA and variable clustering? > Only a little. But anyway - a significant difference should be, that PCA has orthogonal factors and variable clustering's are arbitrarily oblique by its conception.
Gottfried Helms . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
