Hi Steve
> 1)What manager in the mining or petroleum industry
> who has graduated
> from college hasn't taken a serious statistics
> course, including covariances and correlations?
Oh my, you obviously haven't visited many mines. I
can't speak for the petroleum engineers but few miners
(and even fewer geologists) get anything I would
describe as a "serious" statistics course. I have
taught mining engineers at undergraduate and graduate
level on four continents and most of the statistics
they get is taught as a small part of the omineral
resourec evaluation component, mostly by in-department
people with practical experience rather than
statistical qualifications.
> 2) Surely when starting from scratch, educating
> someone about
> geostatistics is more intuitive using covariances?
> (Just my opinion so far,
> speaking as a mathematician who remembers teaching
> basic college level
> statistics to nursing majors, education majors,
> sociology majors, etc. And even succeeding
> occasionally.)
If you have students who have had a good basic
statistics course, then I would agree that a
covariance can form a good basis for the intuitive
development. Most of the people I teach (undergrad,
postgrad and open short courses) have little or no
prior statistical training.
In a 65 hour course, we have time to go through
covariance and least squares regression, but I still
find that the difference approach is easier to get
across than a cross product.
In a 30 hour course, there is little time for basics
and it is far easier to explain something based on
differences than on cross products. The main sticking
point comes when you try to explain the 'diagonal
terms' and why they come out of nowhere. At that
point, covariances can be more useful.
I think one of the confusing factors for most
non-specialist users in when software encourages you
to construct, interpret and model a semi-variogram
function but then insists on a covariance function for
the kriging. In my experience with engineers and
geologists, switching horses in mid-stream leads to
more cold duckings than otherwise. Since the two are
almost always equivalent, why confuse the user?
There is a subsidiary point here, which is not really
pertinent to our discussion, but might be worth
pointing out. Some of the software packages which use
covariances do not go to the full variance at zero
(i.e. the semi-variogram does not go to zero). These
packages go to the equivalent of the nugget effect at
zero and produce rather different answers in many
applications. We discuss this point in our book.
> 3) ........ It seems that
> those instructors did not trust the variogram to be
> more intuitive than spatial covariance functions.
I think this shows the preference of the instructor
and is not a bad thing. If the instructor understands
covariances better than differences, he will teach
better and that, in the final analysis, is what
matters.
> 4) The two basic level introductions to
> geostatistics I have on my
> bookshelf replicate the experience at those two
> classes....
You obviously don't have my 1979 Practical
Geostatistics then. Download it for free at
http://uk.geocities.com/drisobelclark/practica.html
Go with what works for your students on the basis of
the grounding they already have in statistics. I envy
you.
Isobel
____________________________________________________________
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
