[R-sig-Geo] Moran I and General Trend

[Sotos]

Dear All,

I am a bit confused regarding the use of Global Moran
and I hope you can help me clarify the following:

Moran's I, should not be used if the mean and the
variance are not constant in the layer. So, if there
is a trend (first order effect) in the data  (i.e. a
north - south divide) Moran's I should not be used. 

However isn't that a positive autocorrelation in the
data (low with low and high with high)? Isn't that
what Moran is actually measuring?

I understand that in this case local moran is more
proper to use, but what about Global Moran in the
above situation?

Any thoughts?
Thanks in advance
s.




	

	
		
___________________________________________________________ 


διαθέτει την καλύτερη δυνατή προστασία κατά των ενοχλητικών 
μηνυμάτων http://login.yahoo.com/config/mail?.intl=gr