Dear Denis
I am sorry you think that I am being agressive. I
thought I was being quite reasonable, but perception
is a subjective thing. I think it is important for
readers of this list to understand that there are
different ways of coming to the same answer and that
there are different opinions between people in that
field. I was also under the impression that the
purpose of this list is to promote free interchange of
information and opinions.
I have reached a stage in my life where I realise that
other people know much more than I do about many
things, including geostatistics. I joined the list --
as I go to conferences and read journals -- to find
out what other people have to say and to learn new
ways of seeing and doing.
Reading back through this thread of discussion, I am
reminded of the sort of conversations we have at
conferences. A little give-and-take is a valuable
thing provided the parties are actually listening to
one another and hearing what is said, rather than
taking comments personally.
You have every right to prefer LU decomposition and I,
for one, would never deny it to you. It is certainly
more efficient to solve a set of symmetric equations
with a method developed for exactly that purpose. What
I question is the assertion that it is the most
efficient way to tackle the kriging system -
especially if it involves introduction of arbitrary
constants chosen to ensure a positive covariance.
Covariances can be negative and, in the case of an
unbounded semi-variogram, presumably should go
negative at some point.
Since the constant is an artifact and the equations
remain mathematically equivalent to using the
semi-variogram, my original question remains. Why
introduce a complication that is not necessary?
Our opinions on this will always differ. It does not
matter because we still get the same answers in the
end. Surely that is the important fact here?
Isobel Clark
http://uk.geocities.com/drisobelclark
--- Denis ALLARD <[EMAIL PROTECTED]> wrote: >
Dear Isobel,
>
> > Have you not heard of "pivoting"?
>
> I don't understand the point of being so agressive.
> Of course, you are not
> the only person in the geostat community who has
> heard of pivoting. It is
> standard math stuff for any
> mathematician/statistician/engineer.
>
> But what is also standard math stuff is the well
> known fact that if you
> know that your matrix is symetric or def. pos., it
> is much more efficient
> to take advantages of the properties of you matrix.
> So yes, for
> computational efficiency I prefer to use some sort
> of LU decomp for which
> there are much less numerical instabilities if your
> matrix has the higher
> values on the diagonal. Now, if you don't believe
> me, have a look at the
> numerical receipes for instance.
>
> This computational trick is fully transparent for
> the user, and it must
> be so. I don't think that any of us confuses geostat
> and computations.
>
> Now, please try to be less agressive in a forum. Not
> being in full
> agreement with someone is normal in a scientific
> community (and in
> fact in any community), do not take it personaly.
>
> In your previous posts, I have seen a couple of
> assertions that were not
> completely correct. Since this forum is read by many
> students, I believe
> that it is our duty to correct these assertions.
> Now, please do not take
> so agressively. It would much more in your honour to
> acknowledge the
> corrections.
>
> Regards,
>
> Denis Allard
>
>
> PS: about your last question, it is of course silly
> to model the
> covariance function, and I suspect that you know
> that very well....
>
>
> > > It is well known that when inverting a
> > > matrix it is much better (for numerical reasons)
> > > that the higher values
> > > are on the diagonal and the lower values far off
> the
> > > diagonal.
> > Have you not heard of "pivoting"?
> >
> > The computational "problems" of using a matrix
> based
> > on the semi-variogram rather than the covariance
> are
> > removed totally by putting the last equation first
> in
> > the matrix. Of course, this means you cannot use a
> > computational algorithm which demands a symmetric
> > matrix, but so what?
> >
> > Alternatively, you pivot on the largest term in
> the
> > first equation, then the second etc. I did a lot
> of
> > experimentation with this around 15 years ago and
> > found that the two (covariance and semi-variogram)
> > sets of equations become identical after around
> the
> > second or third 'pivot'.
> >
> > Please let us not confuse programming problems
> with
> > geostatistical problems.
> >
> > There are a lot of packages out there which ask
> you to
> > model the semi-variogram and then use a covariance
> for
> > kriging. There are a lot of packages out there
> which
> > model the semi-variogram and krige with the
> > semi-variogram. Are there main stream
> geostatistical
> > (as opposed to statistical or strict GIS) packages
> > which model the covariance?
> >
> > Isobel Clark
> > http://geoecosse.bizland.com
> >
> >
>
____________________________________________________________
> > Do You Yahoo!?
> > Get your free @yahoo.co.uk address at
> http://mail.yahoo.co.uk
> > or your free @yahoo.ie address at
> http://mail.yahoo.ie
> >
>
> .------------------------Denis
> ALLARD--------------------------------.
> | Unite de Biometrie
> [EMAIL PROTECTED] |
> | INRA Domaine St Paul, Site Agroparc tel:
> (33) 4 32 72 21 77 |
> | 84914 AVIGNON cedex 9, FRANCE fax:
> 21 82 |
> |
> |
> `---------
> http://www.avignon.inra.fr/biometrie/welcome.html
> --------'
>
____________________________________________________________
Do You Yahoo!?
Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk
or your free @yahoo.ie address at http://mail.yahoo.ie
--
* To post a message to the list, send it to [EMAIL PROTECTED]
* As a general service to the users, please remember to post a summary of any useful
responses to your questions.
* To unsubscribe, send an email to [EMAIL PROTECTED] with no subject and "unsubscribe
ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND
Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org