Hello, I am new to this forum so apologies if this topic has been raised already. I couldn't find any previous post which actually fully helped with my case. I have circa 270 polygons of different size and shapes for some countries in Asia. A few polygons are dispersed in space, but still have neighbours.
My aim is to run Hot spot analysis (Getis-Ord Gi*) on health related rates. I have run Global Moran's I using different spatial relationships, and I get the highest values of Global Moran's I when using contiguity methods and K-nearest neighbour with 8 neighbours.
All the 3 methods (contiguity edges only, corners, k-nn) result in Global Moran's I index = 0.70 more or less. On what basis do i choose between the three? Are there any pro and cons in using any of the 3 methods with the type of data that i have (poly with different size and shapes)? Should i rather go for a distance based methods?