[R] Efficient way to convert covariance to Euclidian distance matrix

[Jari Oksanen]

Rolf Turner <r.turner <at> auckland.ac.nz> writes:

> 
> On 10/31/13 23:14, Takatsugu Kobayashi wrote:

> >
> > I am struggling to come up with an efficient vectorized way to convert
> > 20Kx20K covariance matrix to a Euclidian distance matrix as a surrogate for
> > dissimilarity matrix. Hopefully I can apply multidimensional scaling for
> > mapping these 20K points (commercial products).
> >

> 
> My suspicion is that with a 20K x 20K covariance matrix:
> 
>      * nothing will work
> 
>      * even if it did, the results would be meaningless numerical noise.
> 
> I.e.  Get real.

FWIW, I have tried NMDS for 19.2 K observations. It worked. The results looked 
sensible (i.e., they were not meaningless numerical noise). It was not fast, 
though, but it can be done. 

Cheers, Jari Oksanen 

PS. Sorry for excessive pruning, but gmane does not allow me post if I don't
remove some quoted lines.