| Re: About REGRESS.pro [message #47794] |
Tue, 07 March 2006 11:27  |
Daelomin
Messages: 7
Registered: March 2006
|
Junior Member |
|
|
Well I know of at least the Numerical Recipes book which has the
regression subroutine which could be tested from a Fortran program...
Is there a thread as to how one can call fortran from IDL? (I'll search
!)
|
|
|
|
|
|
|
|
|
|
|
|
| Re: About REGRESS.pro [message #47877 is a reply to message #47794] |
Wed, 08 March 2006 04:59  |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
Examples are given in the external development guide, available as PDF on
your IDL CD.
Haje
"Daelomin" <joey.pourriciel@gmail.com> wrote in message
news:1141759625.468155.7310@j52g2000cwj.googlegroups.com...
> Well I know of at least the Numerical Recipes book which has the
> regression subroutine which could be tested from a Fortran program...
>
> Is there a thread as to how one can call fortran from IDL? (I'll search
> !)
>
|
|
|
|
| Re: About REGRESS.pro [message #47886 is a reply to message #47795] |
Tue, 07 March 2006 17:53 |
Kenneth P. Bowman
Messages: 585
Registered: May 2000
|
Senior Member |
|
|
In article <MPG.1e777b4ff5bfe993989bc4@news.frii.com>,
David Fanning <davidf@dfanning.com> wrote:
> Kenneth Bowman writes:
>
>> You could test REGRESS by giving it data for which you know the answer.
>
> Too simple. If it was me, and I knew the answer ahead of time,
> confirmation of the result wouldn't make me at all confident.
> I'd just assume they were making the same damn errors I'm
> making. :-(
>
> Cheers,
>
> David
Well, here's a stochastic test. You could also try a problem with an
analytical solution.
Ken
a0 = 1.0
b0 = 0.5
eps = 0.3
n = 1000
x = RANDOMN(seed, n)
y = a0 + b0*x + eps*RANDOMN(seed, n)
b = REGRESS(x, y, CONST = a, YFIT = yfit)
PLOT, x, y, PSYM = 1
OPLOT, x, yfit, PSYM = -3
PRINT, a0, b0
PRINT, a, b
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
