Geography Reference
In-Depth Information
Figure 8.8
Investigating regions of interest in geographic map and spatialization. Break-up of
Pennsylvania into several regions (left) and different geographic regions appearing in relative prox-
imity in the spatialization (right)
such operations onto the geographic map and spatialization. Among such operations are
the various clustering methods commonly applied in multivariate analysis.
In order to demonstrate this, we computed
k
-means clustering on all 200 000
block
groups, based on the very same input vectors used in SOM training. Given the typically
compact nature of
k
-means solutions and the SOM's tendency towards preservation of
neighbourhood relationships, one would expect that clusters computed with the
k
-means
+
Search WWH ::
Custom Search