AI-GEOSTATS: log-normal kriging error

[Marta Rufino]

Dear list members,

I have been calculating log-normal kriging for a data set using geoR 
and I obtained suspicious estimations... so I tried with another data set, 
and I got the same...
I am wondering that maybe there is another problem like the data not 
being strictly log-normal.. could this give such a huge variances?

Sample data is (min, mean and max):
0   23  314
std. error: 59.55
If I do OK, I get:
predictions:
-6  31  218
std err.:
27  40  55
If I do log-OK I get:
predictions:
0.5237  333
std. err.:
2   140 584
Any ideia why I get such a huge variances when using log-normal kriging?
Is it a problem of the data distribution? Is there any possibility of a 
code bug???

Any comment will be mch apreciated!

Thank you
Marta






--
* 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

Re: AI-GEOSTATS: log-normal kriging error

[Isobel Clark]

Marta

I am not familiar with the software you are using, but
it looks like your lognormal standard errors are being
back-transformed into 'raw' units. If this is the case
part of the backtransform is to multiply the 'relative
standard error' by the actual value of the estimate.

That is, if your estimate is 0.52, the backtransformed
standard error is multiplied by 0.52. If the estimate
is 520, it is multiplied by 520. If you do a ratio
between the standard error and the estimate, you will
probably get equivalent 'relative' errors.

We find it better to leave the standard errors in
logarithms and use Sichel type confidence interval,
using the lognormal model. 

Isobel
http://uk.geocities.com/drisobelclark


Download Yahoo! Messenger now for a chance to win Live At Knebworth DVDs
http://www.yahoo.co.uk/robbiewilliams

--
* 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