[R] exponential of a matrix

[Martin Maechler]

>>>>> "MM" == Martin Maechler <maechler at stat.math.ethz.ch>
>>>>>     on Fri, 14 Nov 2008 18:05:54 +0100 writes:

>>>>> "DW" == David Winsemius <dwinsemius at comcast.net>
>>>>>     on Tue, 11 Nov 2008 17:52:18 -0500 writes:

    DW> In 2004 Martyn Plummer said that functionality was
    DW> available in packages msm (as MatrixExp) and Lindsey's
    DW> rmutils (as mexp)

    MM> Well, these two use one of the very dubious ways mentioned below.

oops!  I'm wrong :   MatrixExp() in 'msm',

does, by default (method = "pade") say to use the recommended standard
algorithm (squaring + Pade + scaling).

Last time this came up on the mailing list, for some reason we
had completely overlooked this one ... or it was not yet
available publicly.
My apologies to Chris Jackson, the msm author, whose function
we've overlooked there.

Martin Maechler, ETH Zurich

    DW> David Firth also offered:
    DW> http://www2.warwick.ac.uk/fac/sci/statistics/staff/academic/firth/software/mexp

    MM> yes, and that method (and a slight variation of it which uses
    MM> LAPACK code) is also part of the 'expm' package on R-forge which
    MM> I mentioned in my other reply in this thread (a few minutes
    MM> ago).

    MM> Martin Maechler


    DW> -- David Winsemius


    DW> On Nov 11, 2008, at 5:20 PM, Terry Therneau wrote:

    >>> Is the matrix exponential available in some package?
    >>> 
    >>> The cannonical reference is "Nineteen dubious ways to
    >>> take the exponential of a matrix".  (Love that title)
    >>> 
    >>> Terry T.

    MM> ______________________________________________
    MM> R-help at r-project.org mailing list
    MM> https://stat.ethz.ch/mailman/listinfo/r-help
    MM> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
    MM> and provide commented, minimal, self-contained, reproducible code.