[RsR] Where can I find the source code for cov.rob?

Martin Maechler m@ech|er @end|ng |rom @t@t@m@th@ethz@ch
Fri Mar 13 22:36:06 CET 2009


>>>>> "TT" == Teruko Takada <takada using bus.osaka-cu.ac.jp>
>>>>>     on Fri, 13 Mar 2009 13:10:04 +0900 (JST) writes:

    TT> Dear all,

    TT> Please let me know where I can find all the related
    TT> source codes (possibly R and C) for cov.rob()
    TT> which is the function for minimum volume ellipsoid
    TT> estimation.

    TT> I found that the R function cov.rob() exhibits
    TT> excellent performance in analyzing the tail behavior
    TT> of those data distributed as fat-tailed.
    TT> Therefore, I would like to examine all the detailed
    TT> source codes for the computation of cov.rob().

    TT> After installation of R binary packages,
    TT> the library MASS is located as R/library/MASS
    TT> where all the other packages are placed.
    TT> However, in the R source file packages
    TT> (R-2.8.1.tar.gz and R-latest.tar.gz),
    TT> the library MASS cannot be found under the folder
    TT> R*/src/library, where all the package
    TT> folders are supposed to be placed.

well, it can be found there, but it's wrapped a bit more than
you expected:
The 'Recommended' packages are all added in compressed form,
and MASS is part of the VR bundle 

    TT> In the package VR, 

it's not a package : VR is a "bundle" of 4 packages,
one of which MASS which has the sources you want.

cov.rob() -->  .C(mve_fitlots, ...)
and the C code is in  VR/MASS/src/lqs.c 


    TT> In the package VR, I found MASS.c and cov.trob.R.
    TT> Though cov.rob is explained in the manual of VR,
    TT> I cannot find the source code in the package VR.

    TT> I greatly appreciate your help, in advance.

you're welcome
Martin Maechler, ETH Zurich

    TT> Best regards,

    TT> p.s. This is the first time for me to submit to this
    TT> mailing list. Please excuse me if there is any
    TT> inappropriate things.




More information about the R-SIG-Robust mailing list