Sorry I didn't answer before.
Martin, thanks very much for your notes and for tidying the function
up!= ! I've learnt a lot with you.
About matPower, I didn't realise that where it says "pot <- pot %*% p ot",
consumption of memory was higer than it was on the original
algorithm,= where a function to calculate M^2 where implemented.
And about the other function in this thread, matExp, I can remember I
us= ed Horner's algorithm to get an aproximation to it. I must have a
look to m= y old library.
>Message: 27
>Date: Thu, 22 Jan 2004 18:22:27 +0100
>= From: Martin Maechler <[EMAIL PROTECTED]>
>Subject: [R= ] Re: matrix __power__ (was "exponential")
>To: Vicente Canto Casasol= a <[EMAIL PROTECTED]>
>Cc: [EMAIL PROTECTED] h.ethz.ch
>Message-ID: <[EMAIL PROTECTED] L>
>Content-Type: text/plain; charset=us-ascii
>
>&g= t;>>>> "Vicente" == Vicente Canto Casasola <vicented.can [EMAIL PROTECTED]>
>>>>>> &nbs= p; on Thu, 22 Jan 2004 14:34:01 +0100 writes:
>
> &nb= sp; Vicente> H i, all! First of all, I'd like to apologize fo r
> Vicente> my poor English. It's for years I do= n't use it.
>
>no problem at all.
Well, that is because I'm not speaking!! ;)
>
> Vicente> This is a R-version of a funct= ion I wrote a long
> Vicente> ago for my HP48 cal= culator. It works with the
> Vicente> binar= y expression of the power and just need to
> Vicente= > duplicate the mem used by X.
>
>excellent. This is really = the way to solve the problem
>I think.
>
>As I've menti= oned earlier in this thread,
>computing a matrix "power" is really mu= ch easier than the
>matrix exponential.
I do agree. It did sound strange to me, but I followed the thread!
>
>Hence I wouldn't use exponential in the function name.
&g= t;Also note that trailing ";" are considered as `dirty' (they are
>co= mpletely superfluous).
A C++ tic I have to correct!! Thank you, Martin.
>
>These slight modifications (+ initial "test")
>give>
>matPower <- function(X,n)
>## Function to calculate = the n-th power of a matrix X
>{
> if(n != ro= und(n)) {
> n <- round(n) > warning("rounding exponen= t `n' to", n)
> }
> phi <-= diag(nrow = nrow(X))
> pot <- X # the first po= wer of the matrix.
>
> while (n > 0)
>= ; {
> if (n= %% 2)
> &= nbsp;phi <- phi %*% pot
>
> &nb= sp; n <- n %/% 2
> p= ot <- pot %*% pot
> }
> return(= phi)
>}
______________________________________________
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
