I have used logistic regression to see the correlation
between spatial variables.
As I can not understand well enough what you would like to do,
please forgive me if my answer is stupid.
What I did is;
First get(interpolate) values of independent variables
by interpolation of appropriate kind for each the grid
where the dependent variable is located.
Then export the (y,x1,x2,x3,...)
into ascii file. [ y:dependent variable, x1,x2..:independent variable]
Import the ascii file into such software packages as
STATA, SPLUS, etc.
Now you can do logistic regression.
Contribution of each independent variable is depicted in
forms of odds ratio or p-value.
The fitness of regression model can be
shown as maximum likelihood, hosmer-lemeshow fitness, Wald-test or chi^2.
I am not sure ARC/INFO has such the function of exportation.
I made program of some sort myself.
This mail address is temporary.
International Centre for Medical Research
Kobe University School of Medicine
I am trying to use logistic regression analysis to
predict land cover types particularly cropland in Mountain
region from differtent thematic data ( temp, reainfall etc.)
The cutoff probability of the predicted surface will be used as
a priori probability for maximum likelihood classification of RS
image. I am using ARC/INFO GRID module. My question is how
can I check the contribution of explanatory variables.
I am trying to link it with Bayesian model. How can I do that in
Arc/Info? Is my approach correct? Has anyone used Logistic regression
for predicting land cover types? Could anybody suggest me
any reference in the web resource? Any suugestion please.
*To post a message to the list, send it to ai-geostats@...
*As a general service to list users, please remember to post a summary
of any useful responses to your questions.
*To unsubscribe, send email to majordomo@...
with no subject and
"unsubscribe ai-geostats" in the message body.
DO NOT SEND Subscribe/Unsubscribe requests to the list!