Sorry, an error occurred while loading the content.

## Re: AI-GEOSTATS: Multinomial (Discrete) Kriging

Expand Messages
• are you working with nominal or ordinal data? if you are interested in inferences on the mean and if spatial correlation isn t exceptional, then you can lump
Message 1 of 2 , Nov 29, 2001
are you working with nominal or ordinal data? if you are interested in
inferences on the mean and if spatial correlation isn't exceptional, then
you can lump correlation with other sources of overdispersion and use an
overdispersion estimate to adjust the expected variance (ie
psi*pi(1-pi)/n). since you're interested in the correlation itself you may
be in a different world. within a modeling context, multinomial analysis
often assumes an underlying or latent distribution (distributions plural
under some assumptions). Chiles and Delfiner (pp100-103) demonstrate how
the correlation associated with a latent bivariate Gaussian is more or less
retained when dichotomized near the mean but is rapidly lost as the
dichotomization moves away from the latent mean. I would imagine (but
don't know) that--using the median and shape--similar arguments could be
made for other latent distributions. this would appear to suggest that you
may expect correlation to be highest for nominal or cumulative categories
with prevalence nearest .5. if your data are not at least 1st-order
stationary, then you may expect both the categories nearest the median and
the correlation of all categories to change over space. I would appreciate
other comments on this topic. brian

***************************************************************
Brian Gray
USGS Upper Midwest Environmental Sciences Center
575 Lester Road
Onalaska, WI 54650
608-783-7550 ext 19
FAX 608-783-8058
brgray@...
****************************************************************

Kenneth Cabrera
<krcabrer@... To: "ai-geostats@..." <ai-geostats@...>
ed.edu.co> cc:
Sent by: Subject: AI-GEOSTATS: Multinomial (Discrete) Kriging
ai-geostats-list@unil.c
h

11/29/2001 08:27 AM
Please respond to
Kenneth Cabrera

Hi list!

I want to know if there any software that works with multinomial
(discrete) spatial variable.

Is it valid to work pairwaise with binomial variable and make a join
to obtain the multinomial distribution?

Thank you very much for your help

Kenneth

--
* To post a message to the list, send it to ai-geostats@...
* 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 majordomo@... 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

--
* To post a message to the list, send it to ai-geostats@...
* 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 majordomo@... 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
Your message has been successfully submitted and would be delivered to recipients shortly.