[R] nearest correlation to polychoric

Martin Maechler maechler at stat.math.ethz.ch
Sat Jul 14 10:53:43 CEST 2007


>>>>> "RV" == Ravi Varadhan <rvaradhan at jhmi.edu>
>>>>>     on Fri, 13 Jul 2007 10:52:36 -0400 writes:

    RV> Martin, I sent you the Matlab code for this which I had
    RV> obtained from Nich Higham.

    RV> Cheng, Sheung Hun and Higham, Nick (1998) A Modified
    RV> Cholesky Algorithm Based on a Symmetric Indefinite
    RV> Factorization; \emph{SIAM J. Matrix Anal.\ Appl.},
    RV> \bold{19}, 1097--1110.

    RV> Do you remember?

I now do .. last November.  I'm embarrassed to admit that I hadn't found the
time at the time and had completely forgotten about it.

But now that Jens already has a version of this in R -- thanks
Jens -- this should really find a home in the "R universe"

As a matter of fact, I was thinking of porting the algorithm to
the 'Matrix' package eventually,
the specific "polychor" related case may still better fit to
John Fox' package.

Martin



More information about the R-help mailing list