importance of correlation between data for a PROC CLUSTER

Question

i'm working on a clustering analysis on SAS.

I need to improve an actual code :

PROC CLUSTER METHOD=WARD DATA= Distance_matrix
NOEIGEN
CCC
PSEUDO
PRINT = 50
OUTTREE= Tree; 
id item1;

This is doing a Hierarchical clustering using the ward distance. The data are a distance matrix on product x product.

And i was wondering about the NOEIGEN argument of this method. SAS say :

NOEIGEN: suppresses computation of the eigenvalues of the covariance matrix and substitutes the variances of the variables for the eigenvalues when computing the cubic clustering criterion. The NOEIGEN option saves time if the number of variables is large, but it should be used only if the variables are nearly uncorrelated. If you specify the NOEIGEN option and the variables are highly correlated, the cubic clustering criterion might be very liberal. The NOEIGEN option applies only to coordinate data.

But, automatically there is some correlation in the data no? Otherwise we wouldn't be able to cluster them in categories...

Answer 1

When there are a large number of variables, a whole lot of small correlations can lead to useful clusters. Of course, defining what exactly "large" and "small" and "useful" mean would need a lot of Monte Carlo work.

Note, however, the last sentence of the highlighted text: NOEIGEN only works on coordinate data and you have a distance matrix.