Suppose "Point Moran's I" is defined here as
contribution of each Xi (value at i_th point) to Moran's I.
APIi: Actual, or in other words, observed PI calculated for Xi
Then Xi are permutated randomly many times, say, 1,000 times.
For each permutation, Point Moran's I ( tempPIi) is calculated.
(1)If the above APIi is within lower 5% of tempPIi array( within lower 50 ),
can it be said that Xi is 'spatial outlier'( point of too low/high value)?
(2)If the above APIi is within upper 5% of tempPIi array( within upper 50 ),
can it be said that Xi is in cluster( ' Hot Spot' )?
Or else, point I is just meaningless to locate spatial abnormality?
( I found point I is stable by random permutation. )
By the way, apart from Moran's I, are there any good and simple algorithm
which can 'locate' a spatial abnormality using Random Permutation?
I am afraid( and nearly sure) that my English is failing to
transfer what I would like to say.
Yoshiro Nagao ( Y-Nagao )
International Centre for Medical Research
Kobe University School of Medicine
*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!