I have added the differentiation rules for asin, acos and atan to the file
deriv.c.
Please let me know where to send it.
Date: Thu, 22 Jan 2004 09:05:39 +0100
From: Martin Maechler <maechler at stat.math.ethz.ch>
Subject: Re: [R] derivative of atan(x) and similar functions
To: Timur Elzhov <Timur.Elzhov at jinr.ru>
Cc: R-help <R-help at stat.math.ethz.ch>
Message-ID: <16399.33875.287060.214681 at gargle.gargle.HOWL>
Content-Type: text/plain; charset=us-ascii
>>>>> "Timur" == Timur Elzhov <Timur.Elzhov at jinr.ru>
>>>>> on Wed, 21 Jan 2004 14:54:58 +0300 writes:
Timur> Dear R experts. 'D()' function recognizes some of
Timur> the analitical functions, such as sin, cos, etc. But
Timur> I'd like to take analytical derivatives from asin,
Timur> atan etc. functions. Are there any R packages
Timur> providing that features?
Not that I know of.
We know that it is one deficiency of D() {and deriv() ...} that
it's not user-extendable.
For the asin/acos/atan , I think we would gladly incorporate
them into the current list --- IF someone volunteers to send us
the patch. Since sqrt() and basic arithmetic is already
supported, this is only a matter of diligency..
It's the file <Rsource>/src/main/deriv.c {and mainly (but not
only) the D(SEXP expr, SEXP var) function starting at line 268
that needs to be extended. If know some Lisp, it might be
easier to get started... ;-)
Martin Maechler <maechler at stat.math.ethz.ch>
http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum LEO C16 Leonhardstr. 27
ETH (Federal Inst. Technology) 8092 Zurich SWITZERLAND
phone: x-41-1-632-3408 fax: ...-1228 <><
------------------------------
derivative of atan(x) and similar functions
2 messages · Kasper Kristensen, Martin Maechler
"Kasper" == Kasper Kristensen <kkr at dfu.min.dk>
on Thu, 22 Jan 2004 15:52:32 +0100 writes:
Kasper> I have added the differentiation rules for asin,
Kasper> acos and atan to the file deriv.c. Please let me
Kasper> know where to send it.
Wow, that's a quick response,
thank you very much!
(Kasper has sent the file to me in the mean time).
I'll try to merge it into R-devel (where it deriv.c has been
improved in the initialization part).
Martin